From 4c5827f86d39d12539ec7fead4407bb64ebd6838 Mon Sep 17 00:00:00 2001 From: mrdrivingduck Date: Mon, 16 Oct 2023 11:31:33 +0800 Subject: [PATCH] feat: support smlar extension --- external/Makefile | 1 + external/smlar/.gitignore | 4 + external/smlar/Makefile | 26 + external/smlar/README | 116 ++ external/smlar/expected/bit.out | 105 ++ external/smlar/expected/bytea.out | 149 ++ external/smlar/expected/char.out | 59 + external/smlar/expected/cidr.out | 149 ++ external/smlar/expected/composite_int4.out | 542 ++++++ external/smlar/expected/composite_text.out | 551 ++++++ external/smlar/expected/date.out | 149 ++ external/smlar/expected/float4.out | 149 ++ external/smlar/expected/float8.out | 149 ++ external/smlar/expected/inet.out | 149 ++ external/smlar/expected/int2.out | 149 ++ external/smlar/expected/int4.out | 149 ++ external/smlar/expected/int4g.out | 1633 +++++++++++++++++ external/smlar/expected/int4i.out | 42 + external/smlar/expected/int8.out | 149 ++ external/smlar/expected/int8g.out | 1633 +++++++++++++++++ external/smlar/expected/int8i.out | 42 + external/smlar/expected/interval.out | 149 ++ external/smlar/expected/intervalg.out | 1633 +++++++++++++++++ external/smlar/expected/intervali.out | 42 + external/smlar/expected/macaddr.out | 149 ++ external/smlar/expected/money.out | 105 ++ external/smlar/expected/numeric.out | 149 ++ external/smlar/expected/oid.out | 149 ++ external/smlar/expected/smlar.out | 355 ++++ external/smlar/expected/text.out | 149 ++ external/smlar/expected/textg.out | 1633 +++++++++++++++++ external/smlar/expected/texti.out | 42 + external/smlar/expected/time.out | 149 ++ external/smlar/expected/timestamp.out | 149 ++ external/smlar/expected/timestamptz.out | 149 ++ external/smlar/expected/timetz.out | 149 ++ external/smlar/expected/varbit.out | 105 ++ external/smlar/expected/varchar.out | 149 ++ external/smlar/smlar--1.0.sql | 714 +++++++ external/smlar/smlar--unpackaged--1.0.sql | 125 ++ external/smlar/smlar.c | 1021 +++++++++++ external/smlar/smlar.control | 5 + external/smlar/smlar.h | 144 ++ external/smlar/smlar.sql.in.in | 713 +++++++ external/smlar/smlar_cache.c | 301 +++ external/smlar/smlar_gin.c | 152 ++ external/smlar/smlar_gist.c | 1251 +++++++++++++ external/smlar/smlar_guc.c | 229 +++ external/smlar/smlar_stat.c | 206 +++ external/smlar/sql/bit.sql | 29 + external/smlar/sql/bytea.sql | 34 + external/smlar/sql/char.sql | 31 + external/smlar/sql/cidr.sql | 34 + external/smlar/sql/composite_int4.sql | 46 + external/smlar/sql/composite_text.sql | 46 + external/smlar/sql/date.sql | 34 + external/smlar/sql/float4.sql | 34 + external/smlar/sql/float8.sql | 34 + external/smlar/sql/inet.sql | 34 + external/smlar/sql/int2.sql | 34 + external/smlar/sql/int4.sql | 34 + external/smlar/sql/int4g.sql | 143 ++ external/smlar/sql/int4i.sql | 15 + external/smlar/sql/int8.sql | 34 + external/smlar/sql/int8g.sql | 143 ++ external/smlar/sql/int8i.sql | 15 + external/smlar/sql/int8n.sql | 27 + external/smlar/sql/interval.sql | 34 + external/smlar/sql/intervalg.sql | 143 ++ external/smlar/sql/intervali.sql | 15 + external/smlar/sql/intervaln.sql | 27 + external/smlar/sql/macaddr.sql | 34 + external/smlar/sql/money.sql | 28 + external/smlar/sql/numeric.sql | 34 + external/smlar/sql/oid.sql | 34 + external/smlar/sql/smlar.sql | 136 ++ external/smlar/sql/text.sql | 34 + external/smlar/sql/textg.sql | 143 ++ external/smlar/sql/texti.sql | 15 + external/smlar/sql/textn.sql | 27 + external/smlar/sql/time.sql | 34 + external/smlar/sql/timestamp.sql | 35 + external/smlar/sql/timestamptz.sql | 34 + external/smlar/sql/timetz.sql | 34 + external/smlar/sql/varbit.sql | 28 + external/smlar/sql/varchar.sql | 34 + external/smlar/tsarr.c | 38 + .../translate/CTranslatorScalarToDXL.cpp | 22 +- src/backend/utils/misc/guc.c | 1 + src/include/nodes/primnodes.h | 5 - src/include/utils/px_sync_guc_name.h | 6 + 91 files changed, 18129 insertions(+), 8 deletions(-) create mode 100644 external/smlar/.gitignore create mode 100644 external/smlar/Makefile create mode 100644 external/smlar/README create mode 100644 external/smlar/expected/bit.out create mode 100644 external/smlar/expected/bytea.out create mode 100644 external/smlar/expected/char.out create mode 100644 external/smlar/expected/cidr.out create mode 100644 external/smlar/expected/composite_int4.out create mode 100644 external/smlar/expected/composite_text.out create mode 100644 external/smlar/expected/date.out create mode 100644 external/smlar/expected/float4.out create mode 100644 external/smlar/expected/float8.out create mode 100644 external/smlar/expected/inet.out create mode 100644 external/smlar/expected/int2.out create mode 100644 external/smlar/expected/int4.out create mode 100644 external/smlar/expected/int4g.out create mode 100644 external/smlar/expected/int4i.out create mode 100644 external/smlar/expected/int8.out create mode 100644 external/smlar/expected/int8g.out create mode 100644 external/smlar/expected/int8i.out create mode 100644 external/smlar/expected/interval.out create mode 100644 external/smlar/expected/intervalg.out create mode 100644 external/smlar/expected/intervali.out create mode 100644 external/smlar/expected/macaddr.out create mode 100644 external/smlar/expected/money.out create mode 100644 external/smlar/expected/numeric.out create mode 100644 external/smlar/expected/oid.out create mode 100644 external/smlar/expected/smlar.out create mode 100644 external/smlar/expected/text.out create mode 100644 external/smlar/expected/textg.out create mode 100644 external/smlar/expected/texti.out create mode 100644 external/smlar/expected/time.out create mode 100644 external/smlar/expected/timestamp.out create mode 100644 external/smlar/expected/timestamptz.out create mode 100644 external/smlar/expected/timetz.out create mode 100644 external/smlar/expected/varbit.out create mode 100644 external/smlar/expected/varchar.out create mode 100644 external/smlar/smlar--1.0.sql create mode 100644 external/smlar/smlar--unpackaged--1.0.sql create mode 100644 external/smlar/smlar.c create mode 100644 external/smlar/smlar.control create mode 100644 external/smlar/smlar.h create mode 100644 external/smlar/smlar.sql.in.in create mode 100644 external/smlar/smlar_cache.c create mode 100644 external/smlar/smlar_gin.c create mode 100644 external/smlar/smlar_gist.c create mode 100644 external/smlar/smlar_guc.c create mode 100644 external/smlar/smlar_stat.c create mode 100644 external/smlar/sql/bit.sql create mode 100644 external/smlar/sql/bytea.sql create mode 100644 external/smlar/sql/char.sql create mode 100644 external/smlar/sql/cidr.sql create mode 100644 external/smlar/sql/composite_int4.sql create mode 100644 external/smlar/sql/composite_text.sql create mode 100644 external/smlar/sql/date.sql create mode 100644 external/smlar/sql/float4.sql create mode 100644 external/smlar/sql/float8.sql create mode 100644 external/smlar/sql/inet.sql create mode 100644 external/smlar/sql/int2.sql create mode 100644 external/smlar/sql/int4.sql create mode 100644 external/smlar/sql/int4g.sql create mode 100644 external/smlar/sql/int4i.sql create mode 100644 external/smlar/sql/int8.sql create mode 100644 external/smlar/sql/int8g.sql create mode 100644 external/smlar/sql/int8i.sql create mode 100644 external/smlar/sql/int8n.sql create mode 100644 external/smlar/sql/interval.sql create mode 100644 external/smlar/sql/intervalg.sql create mode 100644 external/smlar/sql/intervali.sql create mode 100644 external/smlar/sql/intervaln.sql create mode 100644 external/smlar/sql/macaddr.sql create mode 100644 external/smlar/sql/money.sql create mode 100644 external/smlar/sql/numeric.sql create mode 100644 external/smlar/sql/oid.sql create mode 100644 external/smlar/sql/smlar.sql create mode 100644 external/smlar/sql/text.sql create mode 100644 external/smlar/sql/textg.sql create mode 100644 external/smlar/sql/texti.sql create mode 100644 external/smlar/sql/textn.sql create mode 100644 external/smlar/sql/time.sql create mode 100644 external/smlar/sql/timestamp.sql create mode 100644 external/smlar/sql/timestamptz.sql create mode 100644 external/smlar/sql/timetz.sql create mode 100644 external/smlar/sql/varbit.sql create mode 100644 external/smlar/sql/varchar.sql create mode 100644 external/smlar/tsarr.c diff --git a/external/Makefile b/external/Makefile index 6ce03f27354..3a235e3b157 100755 --- a/external/Makefile +++ b/external/Makefile @@ -16,6 +16,7 @@ SUBDIRS += polar_parameter_check SUBDIRS += polar_stat_sql SUBDIRS += polar_csn SUBDIRS += polar_px +SUBDIRS += smlar endif #enable_polar_minimal # Common end diff --git a/external/smlar/.gitignore b/external/smlar/.gitignore new file mode 100644 index 00000000000..5dcb3ff9723 --- /dev/null +++ b/external/smlar/.gitignore @@ -0,0 +1,4 @@ +# Generated subdirectories +/log/ +/results/ +/tmp_check/ diff --git a/external/smlar/Makefile b/external/smlar/Makefile new file mode 100644 index 00000000000..d513dc1a967 --- /dev/null +++ b/external/smlar/Makefile @@ -0,0 +1,26 @@ + +MODULE_big = smlar +OBJS = smlar.o smlar_gist.o smlar_gin.o smlar_cache.o \ + tsarr.o smlar_guc.o smlar_stat.o + +EXTENSION = smlar +DATA = smlar--1.0.sql smlar--unpackaged--1.0.sql +REGRESS = smlar int2 int4 int8 float4 float8 money oid \ + timestamp timestamptz time timetz date interval \ + macaddr inet cidr \ + text varchar char bytea bit varbit numeric \ + int4g int8g intervalg textg \ + int4i int8i intervali texti \ + composite_int4 composite_text + +ifdef USE_PGXS +PG_CONFIG = pg_config +PGXS := $(shell $(PG_CONFIG) --pgxs) +include $(PGXS) +else +subdir = external/smlar +top_builddir = ../.. +include $(top_builddir)/src/Makefile.global +include $(top_srcdir)/contrib/contrib-global.mk +endif + diff --git a/external/smlar/README b/external/smlar/README new file mode 100644 index 00000000000..8fa81c51d74 --- /dev/null +++ b/external/smlar/README @@ -0,0 +1,116 @@ +float4 smlar(anyarray, anyarray) + - computes similary of two arrays. Arrays should be the same type. + +float4 smlar(anyarray, anyarray, bool useIntersect) + - computes similary of two arrays of composite types. Composite type looks like: + CREATE TYPE type_name AS (element_name anytype, weight_name FLOAT4); + useIntersect option points to use only intersected elements in denominator + see an exmaples in sql/composite_int4.sql or sql/composite_text.sql + +float4 smlar( anyarray a, anyarray b, text formula ); + - computes similary of two arrays by given formula, arrays should + be the same type. + Predefined variables in formula: + N.i - number of common elements in both array (intersection) + N.a - number of uniqueelements in first array + N.b - number of uniqueelements in second array + Example: + smlar('{1,4,6}'::int[], '{5,4,6}' ) + smlar('{1,4,6}'::int[], '{5,4,6}', 'N.i / sqrt(N.a * N.b)' ) + That calls are equivalent. + +anyarray % anyarray + - returns true if similarity of that arrays is greater than limit + +float4 show_smlar_limit() - deprecated + - shows the limit for % operation + +float4 set_smlar_limit(float4) - deprecated + - sets the limit for % operation + +Use instead of show_smlar_limit/set_smlar_limit GUC variable +smlar.threshold (see below) + + +text[] tsvector2textarray(tsvector) + - transforms tsvector type to text array + +anyarray array_unique(anyarray) + - sort and unique array + +float4 inarray(anyarray, anyelement) + - returns zero if second argument does not present in a first one + and 1.0 in opposite case + +float4 inarray(anyarray, anyelement, float4, float4) + - returns fourth argument if second argument does not present in + a first one and third argument in opposite case + +GUC configuration variables: + +smlar.threshold FLOAT + Array's with similarity lower than threshold are not similar + by % operation + +smlar.persistent_cache BOOL + Cache of global stat is stored in transaction-independent memory + +smlar.type STRING + Type of similarity formula: cosine(default), tfidf, overlap + +smlar.stattable STRING + Name of table stored set-wide statistic. Table should be + defined as + CREATE TABLE table_name ( + value data_type UNIQUE, + ndoc int4 (or bigint) NOT NULL CHECK (ndoc>0) + ); + And row with null value means total number of documents. + See an examples in sql/*g.sql files + Note: used on for smlar.type = 'tfidf' + +smlar.tf_method STRING + Calculation method for term frequency. Values: + "n" - simple counting of entries (default) + "log" - 1 + log(n) + "const" - TF is equal to 1 + Note: used on for smlar.type = 'tfidf' + +smlar.idf_plus_one BOOL + If false (default), calculate idf as log(d/df), + if true - as log(1+d/df) + Note: used on for smlar.type = 'tfidf' + +Module provides several GUC variables smlar.threshold, it's highly +recommended to add to postgesql.conf: +custom_variable_classes = 'smlar' # list of custom variable class names +smlar.threshold = 0.6 #or any other value > 0 and < 1 +and other smlar.* variables + +GiST/GIN support for % and && operations for: + Array Type | GIN operator class | GiST operator class +---------------+----------------------+---------------------- + bit[] | _bit_sml_ops | + bytea[] | _bytea_sml_ops | _bytea_sml_ops + char[] | _char_sml_ops | _char_sml_ops + cidr[] | _cidr_sml_ops | _cidr_sml_ops + date[] | _date_sml_ops | _date_sml_ops + float4[] | _float4_sml_ops | _float4_sml_ops + float8[] | _float8_sml_ops | _float8_sml_ops + inet[] | _inet_sml_ops | _inet_sml_ops + int2[] | _int2_sml_ops | _int2_sml_ops + int4[] | _int4_sml_ops | _int4_sml_ops + int8[] | _int8_sml_ops | _int8_sml_ops + interval[] | _interval_sml_ops | _interval_sml_ops + macaddr[] | _macaddr_sml_ops | _macaddr_sml_ops + money[] | _money_sml_ops | + numeric[] | _numeric_sml_ops | _numeric_sml_ops + oid[] | _oid_sml_ops | _oid_sml_ops + text[] | _text_sml_ops | _text_sml_ops + time[] | _time_sml_ops | _time_sml_ops + timestamp[] | _timestamp_sml_ops | _timestamp_sml_ops + timestamptz[] | _timestamptz_sml_ops | _timestamptz_sml_ops + timetz[] | _timetz_sml_ops | _timetz_sml_ops + varbit[] | _varbit_sml_ops | + varchar[] | _varchar_sml_ops | _varchar_sml_ops + diff --git a/external/smlar/expected/bit.out b/external/smlar/expected/bit.out new file mode 100644 index 00000000000..fff8c6e2909 --- /dev/null +++ b/external/smlar/expected/bit.out @@ -0,0 +1,105 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::int4::bit(10) + FROM + generate_series(1, t) as v + ) AS v + INTO test_bit +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[]) AS s FROM test_bit WHERE v % '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[]) AS s FROM test_bit WHERE v % '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_bit ON test_bit USING gin (v _bit_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[]) AS s FROM test_bit WHERE v % '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[]) AS s FROM test_bit WHERE v % '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/bytea.out b/external/smlar/expected/bytea.out new file mode 100644 index 00000000000..2c3d4c953fb --- /dev/null +++ b/external/smlar/expected/bytea.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::text::bytea + FROM + generate_series(1, t) as v + ) AS v + INTO test_bytea +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_bytea WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_bytea WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_bytea ON test_bytea USING gist (v _bytea_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_bytea WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_bytea WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_bytea; +CREATE INDEX idx_test_bytea ON test_bytea USING gin (v _bytea_sml_ops); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_bytea WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_bytea WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/char.out b/external/smlar/expected/char.out new file mode 100644 index 00000000000..af7e037a45e --- /dev/null +++ b/external/smlar/expected/char.out @@ -0,0 +1,59 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::varchar(1)::"char" + FROM + generate_series(1, t) as v + ) AS v + INTO test_char +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{3,2,1}') AS s FROM test_char WHERE v % '{3,2,1}' ORDER BY s DESC, t; + t | s +---+---------- + 3 | 1 + 4 | 0.866025 + 2 | 0.816497 + 5 | 0.774597 + 6 | 0.707107 + 7 | 0.654654 + 8 | 0.612372 +(7 rows) + +CREATE INDEX idx_test_char ON test_char USING gist (v _char_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{3,2,1}') AS s FROM test_char WHERE v % '{3,2,1}' ORDER BY s DESC, t; + t | s +---+---------- + 3 | 1 + 4 | 0.866025 + 2 | 0.816497 + 5 | 0.774597 + 6 | 0.707107 + 7 | 0.654654 + 8 | 0.612372 +(7 rows) + +DROP INDEX idx_test_char; +CREATE INDEX idx_test_char ON test_char USING gin (v _char_sml_ops); +SELECT t, smlar(v, '{3,2,1}') AS s FROM test_char WHERE v % '{3,2,1}' ORDER BY s DESC, t; + t | s +---+---------- + 3 | 1 + 4 | 0.866025 + 2 | 0.816497 + 5 | 0.774597 + 6 | 0.707107 + 7 | 0.654654 + 8 | 0.612372 +(7 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/cidr.out b/external/smlar/expected/cidr.out new file mode 100644 index 00000000000..031321076b1 --- /dev/null +++ b/external/smlar/expected/cidr.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + ( '192.168.1.' || (v % 256) )::cidr + FROM + generate_series(1, t) as v + ) AS v + INTO test_cidr +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}') AS s FROM test_cidr WHERE v % '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}') AS s FROM test_cidr WHERE v % '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_cidr ON test_cidr USING gist (v _cidr_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}') AS s FROM test_cidr WHERE v % '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}') AS s FROM test_cidr WHERE v % '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_cidr; +CREATE INDEX idx_test_cidr ON test_cidr USING gin (v _cidr_sml_ops); +SELECT t, smlar(v, '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}') AS s FROM test_cidr WHERE v % '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}') AS s FROM test_cidr WHERE v % '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/composite_int4.out b/external/smlar/expected/composite_int4.out new file mode 100644 index 00000000000..1a46c08a91d --- /dev/null +++ b/external/smlar/expected/composite_int4.out @@ -0,0 +1,542 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +CREATE TYPE cint AS (id int, w float4); +SELECT + t, + ARRAY( + SELECT + (ROW(v::int4, v::float4))::cint + FROM + generate_series(1, t) as v + ) AS v + INTO test_composite_int4 +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[]) AS s + FROM test_composite_int4 + WHERE v % '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[] AS s + FROM test_composite_int4 + WHERE v % '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[]) ORDER BY s DESC, t; +ERROR: syntax error at or near "AS" +LINE 1: ..."(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[] AS s + ^ +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[]) AS s + FROM test_composite_int4 + WHERE smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[]) > 0.6 ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[]) AS s + FROM test_composite_int4 + WHERE smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[]) > 0.6 ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[], false) AS s + FROM test_composite_int4 + WHERE smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[], false) > 0.6 ORDER BY s DESC, t; + t | s +----+---------- + 10 | 0.886405 + 9 | 0.842927 + 8 | 0.797053 + 11 | 0.773192 + 7 | 0.748331 + 6 | 0.696143 + 12 | 0.682191 + 5 | 0.639602 + 13 | 0.607744 +(9 rows) + +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[], false) AS s + FROM test_composite_int4 + WHERE smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[], false) > 0.6 ORDER BY s DESC, t; + t | s +---+---------- + 8 | 0.730632 + 7 | 0.668153 + 9 | 0.618147 +(3 rows) + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[], true) AS s + FROM test_composite_int4 + WHERE smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[], true) > 0.6 ORDER BY s DESC, t; + t | s +-----+---------- + 1 | 1 + 2 | 0.948683 + 3 | 0.92582 + 4 | 0.912871 + 5 | 0.904534 + 6 | 0.898717 + 7 | 0.894427 + 8 | 0.891133 + 9 | 0.888523 + 10 | 0.886405 + 11 | 0.886405 + 12 | 0.886405 + 13 | 0.886405 + 14 | 0.886405 + 15 | 0.886405 + 16 | 0.886405 + 17 | 0.886405 + 18 | 0.886405 + 19 | 0.886405 + 20 | 0.886405 + 21 | 0.886405 + 22 | 0.886405 + 23 | 0.886405 + 24 | 0.886405 + 25 | 0.886405 + 26 | 0.886405 + 27 | 0.886405 + 28 | 0.886405 + 29 | 0.886405 + 30 | 0.886405 + 31 | 0.886405 + 32 | 0.886405 + 33 | 0.886405 + 34 | 0.886405 + 35 | 0.886405 + 36 | 0.886405 + 37 | 0.886405 + 38 | 0.886405 + 39 | 0.886405 + 40 | 0.886405 + 41 | 0.886405 + 42 | 0.886405 + 43 | 0.886405 + 44 | 0.886405 + 45 | 0.886405 + 46 | 0.886405 + 47 | 0.886405 + 48 | 0.886405 + 49 | 0.886405 + 50 | 0.886405 + 51 | 0.886405 + 52 | 0.886405 + 53 | 0.886405 + 54 | 0.886405 + 55 | 0.886405 + 56 | 0.886405 + 57 | 0.886405 + 58 | 0.886405 + 59 | 0.886405 + 60 | 0.886405 + 61 | 0.886405 + 62 | 0.886405 + 63 | 0.886405 + 64 | 0.886405 + 65 | 0.886405 + 66 | 0.886405 + 67 | 0.886405 + 68 | 0.886405 + 69 | 0.886405 + 70 | 0.886405 + 71 | 0.886405 + 72 | 0.886405 + 73 | 0.886405 + 74 | 0.886405 + 75 | 0.886405 + 76 | 0.886405 + 77 | 0.886405 + 78 | 0.886405 + 79 | 0.886405 + 80 | 0.886405 + 81 | 0.886405 + 82 | 0.886405 + 83 | 0.886405 + 84 | 0.886405 + 85 | 0.886405 + 86 | 0.886405 + 87 | 0.886405 + 88 | 0.886405 + 89 | 0.886405 + 90 | 0.886405 + 91 | 0.886405 + 92 | 0.886405 + 93 | 0.886405 + 94 | 0.886405 + 95 | 0.886405 + 96 | 0.886405 + 97 | 0.886405 + 98 | 0.886405 + 99 | 0.886405 + 100 | 0.886405 + 101 | 0.886405 + 102 | 0.886405 + 103 | 0.886405 + 104 | 0.886405 + 105 | 0.886405 + 106 | 0.886405 + 107 | 0.886405 + 108 | 0.886405 + 109 | 0.886405 + 110 | 0.886405 + 111 | 0.886405 + 112 | 0.886405 + 113 | 0.886405 + 114 | 0.886405 + 115 | 0.886405 + 116 | 0.886405 + 117 | 0.886405 + 118 | 0.886405 + 119 | 0.886405 + 120 | 0.886405 + 121 | 0.886405 + 122 | 0.886405 + 123 | 0.886405 + 124 | 0.886405 + 125 | 0.886405 + 126 | 0.886405 + 127 | 0.886405 + 128 | 0.886405 + 129 | 0.886405 + 130 | 0.886405 + 131 | 0.886405 + 132 | 0.886405 + 133 | 0.886405 + 134 | 0.886405 + 135 | 0.886405 + 136 | 0.886405 + 137 | 0.886405 + 138 | 0.886405 + 139 | 0.886405 + 140 | 0.886405 + 141 | 0.886405 + 142 | 0.886405 + 143 | 0.886405 + 144 | 0.886405 + 145 | 0.886405 + 146 | 0.886405 + 147 | 0.886405 + 148 | 0.886405 + 149 | 0.886405 + 150 | 0.886405 + 151 | 0.886405 + 152 | 0.886405 + 153 | 0.886405 + 154 | 0.886405 + 155 | 0.886405 + 156 | 0.886405 + 157 | 0.886405 + 158 | 0.886405 + 159 | 0.886405 + 160 | 0.886405 + 161 | 0.886405 + 162 | 0.886405 + 163 | 0.886405 + 164 | 0.886405 + 165 | 0.886405 + 166 | 0.886405 + 167 | 0.886405 + 168 | 0.886405 + 169 | 0.886405 + 170 | 0.886405 + 171 | 0.886405 + 172 | 0.886405 + 173 | 0.886405 + 174 | 0.886405 + 175 | 0.886405 + 176 | 0.886405 + 177 | 0.886405 + 178 | 0.886405 + 179 | 0.886405 + 180 | 0.886405 + 181 | 0.886405 + 182 | 0.886405 + 183 | 0.886405 + 184 | 0.886405 + 185 | 0.886405 + 186 | 0.886405 + 187 | 0.886405 + 188 | 0.886405 + 189 | 0.886405 + 190 | 0.886405 + 191 | 0.886405 + 192 | 0.886405 + 193 | 0.886405 + 194 | 0.886405 + 195 | 0.886405 + 196 | 0.886405 + 197 | 0.886405 + 198 | 0.886405 + 199 | 0.886405 + 200 | 0.886405 +(200 rows) + +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[], true) AS s + FROM test_composite_int4 + WHERE smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[], true) > 0.6 ORDER BY s DESC, t; + t | s +-----+---------- + 1 | 1 + 2 | 0.948683 + 3 | 0.948683 + 8 | 0.893198 + 9 | 0.893198 + 10 | 0.893198 + 11 | 0.893198 + 12 | 0.893198 + 13 | 0.893198 + 14 | 0.893198 + 15 | 0.893198 + 16 | 0.893198 + 17 | 0.893198 + 18 | 0.893198 + 19 | 0.893198 + 20 | 0.893198 + 21 | 0.893198 + 22 | 0.893198 + 23 | 0.893198 + 24 | 0.893198 + 25 | 0.893198 + 26 | 0.893198 + 27 | 0.893198 + 28 | 0.893198 + 29 | 0.893198 + 30 | 0.893198 + 31 | 0.893198 + 32 | 0.893198 + 7 | 0.89172 + 6 | 0.888957 + 5 | 0.884652 + 4 | 0.881917 + 50 | 0.663459 + 51 | 0.663459 + 52 | 0.663459 + 53 | 0.663459 + 54 | 0.663459 + 55 | 0.663459 + 56 | 0.663459 + 57 | 0.663459 + 58 | 0.663459 + 59 | 0.663459 + 60 | 0.663459 + 61 | 0.663459 + 62 | 0.663459 + 63 | 0.663459 + 64 | 0.663459 + 65 | 0.663459 + 66 | 0.663459 + 67 | 0.663459 + 68 | 0.663459 + 69 | 0.663459 + 70 | 0.663459 + 71 | 0.663459 + 72 | 0.663459 + 73 | 0.663459 + 74 | 0.663459 + 75 | 0.663459 + 76 | 0.663459 + 77 | 0.663459 + 78 | 0.663459 + 79 | 0.663459 + 80 | 0.663459 + 81 | 0.663459 + 82 | 0.663459 + 83 | 0.663459 + 84 | 0.663459 + 85 | 0.663459 + 86 | 0.663459 + 87 | 0.663459 + 88 | 0.663459 + 89 | 0.663459 + 90 | 0.663459 + 91 | 0.663459 + 92 | 0.663459 + 93 | 0.663459 + 94 | 0.663459 + 95 | 0.663459 + 96 | 0.663459 + 97 | 0.663459 + 98 | 0.663459 + 99 | 0.663459 + 100 | 0.663459 + 101 | 0.663459 + 102 | 0.663459 + 103 | 0.663459 + 104 | 0.663459 + 105 | 0.663459 + 106 | 0.663459 + 107 | 0.663459 + 108 | 0.663459 + 109 | 0.663459 + 110 | 0.663459 + 111 | 0.663459 + 112 | 0.663459 + 113 | 0.663459 + 114 | 0.663459 + 115 | 0.663459 + 116 | 0.663459 + 117 | 0.663459 + 118 | 0.663459 + 119 | 0.663459 + 120 | 0.663459 + 121 | 0.663459 + 122 | 0.663459 + 123 | 0.663459 + 124 | 0.663459 + 125 | 0.663459 + 126 | 0.663459 + 127 | 0.663459 + 128 | 0.663459 + 129 | 0.663459 + 130 | 0.663459 + 131 | 0.663459 + 132 | 0.663459 + 133 | 0.663459 + 134 | 0.663459 + 135 | 0.663459 + 136 | 0.663459 + 137 | 0.663459 + 138 | 0.663459 + 139 | 0.663459 + 140 | 0.663459 + 141 | 0.663459 + 142 | 0.663459 + 143 | 0.663459 + 144 | 0.663459 + 145 | 0.663459 + 146 | 0.663459 + 147 | 0.663459 + 148 | 0.663459 + 149 | 0.663459 + 150 | 0.663459 + 151 | 0.663459 + 152 | 0.663459 + 153 | 0.663459 + 154 | 0.663459 + 155 | 0.663459 + 156 | 0.663459 + 157 | 0.663459 + 158 | 0.663459 + 159 | 0.663459 + 160 | 0.663459 + 161 | 0.663459 + 162 | 0.663459 + 163 | 0.663459 + 164 | 0.663459 + 165 | 0.663459 + 166 | 0.663459 + 167 | 0.663459 + 168 | 0.663459 + 169 | 0.663459 + 170 | 0.663459 + 171 | 0.663459 + 172 | 0.663459 + 173 | 0.663459 + 174 | 0.663459 + 175 | 0.663459 + 176 | 0.663459 + 177 | 0.663459 + 178 | 0.663459 + 179 | 0.663459 + 180 | 0.663459 + 181 | 0.663459 + 182 | 0.663459 + 183 | 0.663459 + 184 | 0.663459 + 185 | 0.663459 + 186 | 0.663459 + 187 | 0.663459 + 188 | 0.663459 + 189 | 0.663459 + 190 | 0.663459 + 191 | 0.663459 + 192 | 0.663459 + 193 | 0.663459 + 194 | 0.663459 + 195 | 0.663459 + 196 | 0.663459 + 197 | 0.663459 + 198 | 0.663459 + 199 | 0.663459 + 200 | 0.663459 + 33 | 0.651203 + 34 | 0.651203 + 35 | 0.651203 + 36 | 0.651203 + 37 | 0.651203 + 38 | 0.651203 + 39 | 0.651203 + 40 | 0.651203 + 41 | 0.651203 + 42 | 0.651203 + 43 | 0.651203 + 44 | 0.651203 + 45 | 0.651203 + 46 | 0.651203 + 47 | 0.651203 + 48 | 0.651203 + 49 | 0.631477 +(200 rows) + diff --git a/external/smlar/expected/composite_text.out b/external/smlar/expected/composite_text.out new file mode 100644 index 00000000000..48249348d42 --- /dev/null +++ b/external/smlar/expected/composite_text.out @@ -0,0 +1,551 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +CREATE TYPE ctext AS (id text, w float4); +SELECT + t, + ARRAY( + SELECT + (ROW(v::text, v::float4))::ctext + FROM + generate_series(1, t) as v + ) AS v + INTO test_composite_text +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[]) AS s + FROM test_composite_text + WHERE v % '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[]) AS s + FROM test_composite_text + WHERE v % '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[]) AS s + FROM test_composite_text + WHERE smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[]) > 0.6 ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[]) AS s + FROM test_composite_text + WHERE smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[]) > 0.6 ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[], false) AS s + FROM test_composite_text + WHERE smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[], false) > 0.6 ORDER BY s DESC, t; + t | s +----+---------- + 10 | 0.886405 + 9 | 0.842927 + 8 | 0.797053 + 11 | 0.773192 + 7 | 0.748331 + 6 | 0.696143 + 12 | 0.682191 + 5 | 0.639602 + 13 | 0.607744 +(9 rows) + +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[], false) AS s + FROM test_composite_text + WHERE smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[], false) > 0.6 ORDER BY s DESC, t; + t | s +---+---------- + 8 | 0.730632 + 7 | 0.668153 + 9 | 0.618147 +(3 rows) + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[], true) AS s + FROM test_composite_text + WHERE smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[], true) > 0.6 ORDER BY s DESC, t; + t | s +-----+---------- + 1 | 1 + 2 | 0.948683 + 3 | 0.92582 + 4 | 0.912871 + 5 | 0.904534 + 6 | 0.898717 + 7 | 0.894427 + 8 | 0.891133 + 9 | 0.888523 + 10 | 0.886405 + 11 | 0.886405 + 12 | 0.886405 + 13 | 0.886405 + 14 | 0.886405 + 15 | 0.886405 + 16 | 0.886405 + 17 | 0.886405 + 18 | 0.886405 + 19 | 0.886405 + 20 | 0.886405 + 21 | 0.886405 + 22 | 0.886405 + 23 | 0.886405 + 24 | 0.886405 + 25 | 0.886405 + 26 | 0.886405 + 27 | 0.886405 + 28 | 0.886405 + 29 | 0.886405 + 30 | 0.886405 + 31 | 0.886405 + 32 | 0.886405 + 33 | 0.886405 + 34 | 0.886405 + 35 | 0.886405 + 36 | 0.886405 + 37 | 0.886405 + 38 | 0.886405 + 39 | 0.886405 + 40 | 0.886405 + 41 | 0.886405 + 42 | 0.886405 + 43 | 0.886405 + 44 | 0.886405 + 45 | 0.886405 + 46 | 0.886405 + 47 | 0.886405 + 48 | 0.886405 + 49 | 0.886405 + 50 | 0.886405 + 51 | 0.886405 + 52 | 0.886405 + 53 | 0.886405 + 54 | 0.886405 + 55 | 0.886405 + 56 | 0.886405 + 57 | 0.886405 + 58 | 0.886405 + 59 | 0.886405 + 60 | 0.886405 + 61 | 0.886405 + 62 | 0.886405 + 63 | 0.886405 + 64 | 0.886405 + 65 | 0.886405 + 66 | 0.886405 + 67 | 0.886405 + 68 | 0.886405 + 69 | 0.886405 + 70 | 0.886405 + 71 | 0.886405 + 72 | 0.886405 + 73 | 0.886405 + 74 | 0.886405 + 75 | 0.886405 + 76 | 0.886405 + 77 | 0.886405 + 78 | 0.886405 + 79 | 0.886405 + 80 | 0.886405 + 81 | 0.886405 + 82 | 0.886405 + 83 | 0.886405 + 84 | 0.886405 + 85 | 0.886405 + 86 | 0.886405 + 87 | 0.886405 + 88 | 0.886405 + 89 | 0.886405 + 90 | 0.886405 + 91 | 0.886405 + 92 | 0.886405 + 93 | 0.886405 + 94 | 0.886405 + 95 | 0.886405 + 96 | 0.886405 + 97 | 0.886405 + 98 | 0.886405 + 99 | 0.886405 + 100 | 0.886405 + 101 | 0.886405 + 102 | 0.886405 + 103 | 0.886405 + 104 | 0.886405 + 105 | 0.886405 + 106 | 0.886405 + 107 | 0.886405 + 108 | 0.886405 + 109 | 0.886405 + 110 | 0.886405 + 111 | 0.886405 + 112 | 0.886405 + 113 | 0.886405 + 114 | 0.886405 + 115 | 0.886405 + 116 | 0.886405 + 117 | 0.886405 + 118 | 0.886405 + 119 | 0.886405 + 120 | 0.886405 + 121 | 0.886405 + 122 | 0.886405 + 123 | 0.886405 + 124 | 0.886405 + 125 | 0.886405 + 126 | 0.886405 + 127 | 0.886405 + 128 | 0.886405 + 129 | 0.886405 + 130 | 0.886405 + 131 | 0.886405 + 132 | 0.886405 + 133 | 0.886405 + 134 | 0.886405 + 135 | 0.886405 + 136 | 0.886405 + 137 | 0.886405 + 138 | 0.886405 + 139 | 0.886405 + 140 | 0.886405 + 141 | 0.886405 + 142 | 0.886405 + 143 | 0.886405 + 144 | 0.886405 + 145 | 0.886405 + 146 | 0.886405 + 147 | 0.886405 + 148 | 0.886405 + 149 | 0.886405 + 150 | 0.886405 + 151 | 0.886405 + 152 | 0.886405 + 153 | 0.886405 + 154 | 0.886405 + 155 | 0.886405 + 156 | 0.886405 + 157 | 0.886405 + 158 | 0.886405 + 159 | 0.886405 + 160 | 0.886405 + 161 | 0.886405 + 162 | 0.886405 + 163 | 0.886405 + 164 | 0.886405 + 165 | 0.886405 + 166 | 0.886405 + 167 | 0.886405 + 168 | 0.886405 + 169 | 0.886405 + 170 | 0.886405 + 171 | 0.886405 + 172 | 0.886405 + 173 | 0.886405 + 174 | 0.886405 + 175 | 0.886405 + 176 | 0.886405 + 177 | 0.886405 + 178 | 0.886405 + 179 | 0.886405 + 180 | 0.886405 + 181 | 0.886405 + 182 | 0.886405 + 183 | 0.886405 + 184 | 0.886405 + 185 | 0.886405 + 186 | 0.886405 + 187 | 0.886405 + 188 | 0.886405 + 189 | 0.886405 + 190 | 0.886405 + 191 | 0.886405 + 192 | 0.886405 + 193 | 0.886405 + 194 | 0.886405 + 195 | 0.886405 + 196 | 0.886405 + 197 | 0.886405 + 198 | 0.886405 + 199 | 0.886405 + 200 | 0.886405 +(200 rows) + +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[], true) AS s + FROM test_composite_text + WHERE smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[], true) > 0.6 ORDER BY s DESC, t; + t | s +-----+---------- + 1 | 1 + 2 | 0.948683 + 3 | 0.948683 + 8 | 0.893198 + 9 | 0.893198 + 10 | 0.893198 + 11 | 0.893198 + 12 | 0.893198 + 13 | 0.893198 + 14 | 0.893198 + 15 | 0.893198 + 16 | 0.893198 + 17 | 0.893198 + 18 | 0.893198 + 19 | 0.893198 + 20 | 0.893198 + 21 | 0.893198 + 22 | 0.893198 + 23 | 0.893198 + 24 | 0.893198 + 25 | 0.893198 + 26 | 0.893198 + 27 | 0.893198 + 28 | 0.893198 + 29 | 0.893198 + 30 | 0.893198 + 31 | 0.893198 + 32 | 0.893198 + 7 | 0.89172 + 6 | 0.888957 + 5 | 0.884652 + 4 | 0.881917 + 50 | 0.663459 + 51 | 0.663459 + 52 | 0.663459 + 53 | 0.663459 + 54 | 0.663459 + 55 | 0.663459 + 56 | 0.663459 + 57 | 0.663459 + 58 | 0.663459 + 59 | 0.663459 + 60 | 0.663459 + 61 | 0.663459 + 62 | 0.663459 + 63 | 0.663459 + 64 | 0.663459 + 65 | 0.663459 + 66 | 0.663459 + 67 | 0.663459 + 68 | 0.663459 + 69 | 0.663459 + 70 | 0.663459 + 71 | 0.663459 + 72 | 0.663459 + 73 | 0.663459 + 74 | 0.663459 + 75 | 0.663459 + 76 | 0.663459 + 77 | 0.663459 + 78 | 0.663459 + 79 | 0.663459 + 80 | 0.663459 + 81 | 0.663459 + 82 | 0.663459 + 83 | 0.663459 + 84 | 0.663459 + 85 | 0.663459 + 86 | 0.663459 + 87 | 0.663459 + 88 | 0.663459 + 89 | 0.663459 + 90 | 0.663459 + 91 | 0.663459 + 92 | 0.663459 + 93 | 0.663459 + 94 | 0.663459 + 95 | 0.663459 + 96 | 0.663459 + 97 | 0.663459 + 98 | 0.663459 + 99 | 0.663459 + 100 | 0.663459 + 101 | 0.663459 + 102 | 0.663459 + 103 | 0.663459 + 104 | 0.663459 + 105 | 0.663459 + 106 | 0.663459 + 107 | 0.663459 + 108 | 0.663459 + 109 | 0.663459 + 110 | 0.663459 + 111 | 0.663459 + 112 | 0.663459 + 113 | 0.663459 + 114 | 0.663459 + 115 | 0.663459 + 116 | 0.663459 + 117 | 0.663459 + 118 | 0.663459 + 119 | 0.663459 + 120 | 0.663459 + 121 | 0.663459 + 122 | 0.663459 + 123 | 0.663459 + 124 | 0.663459 + 125 | 0.663459 + 126 | 0.663459 + 127 | 0.663459 + 128 | 0.663459 + 129 | 0.663459 + 130 | 0.663459 + 131 | 0.663459 + 132 | 0.663459 + 133 | 0.663459 + 134 | 0.663459 + 135 | 0.663459 + 136 | 0.663459 + 137 | 0.663459 + 138 | 0.663459 + 139 | 0.663459 + 140 | 0.663459 + 141 | 0.663459 + 142 | 0.663459 + 143 | 0.663459 + 144 | 0.663459 + 145 | 0.663459 + 146 | 0.663459 + 147 | 0.663459 + 148 | 0.663459 + 149 | 0.663459 + 150 | 0.663459 + 151 | 0.663459 + 152 | 0.663459 + 153 | 0.663459 + 154 | 0.663459 + 155 | 0.663459 + 156 | 0.663459 + 157 | 0.663459 + 158 | 0.663459 + 159 | 0.663459 + 160 | 0.663459 + 161 | 0.663459 + 162 | 0.663459 + 163 | 0.663459 + 164 | 0.663459 + 165 | 0.663459 + 166 | 0.663459 + 167 | 0.663459 + 168 | 0.663459 + 169 | 0.663459 + 170 | 0.663459 + 171 | 0.663459 + 172 | 0.663459 + 173 | 0.663459 + 174 | 0.663459 + 175 | 0.663459 + 176 | 0.663459 + 177 | 0.663459 + 178 | 0.663459 + 179 | 0.663459 + 180 | 0.663459 + 181 | 0.663459 + 182 | 0.663459 + 183 | 0.663459 + 184 | 0.663459 + 185 | 0.663459 + 186 | 0.663459 + 187 | 0.663459 + 188 | 0.663459 + 189 | 0.663459 + 190 | 0.663459 + 191 | 0.663459 + 192 | 0.663459 + 193 | 0.663459 + 194 | 0.663459 + 195 | 0.663459 + 196 | 0.663459 + 197 | 0.663459 + 198 | 0.663459 + 199 | 0.663459 + 200 | 0.663459 + 33 | 0.651203 + 34 | 0.651203 + 35 | 0.651203 + 36 | 0.651203 + 37 | 0.651203 + 38 | 0.651203 + 39 | 0.651203 + 40 | 0.651203 + 41 | 0.651203 + 42 | 0.651203 + 43 | 0.651203 + 44 | 0.651203 + 45 | 0.651203 + 46 | 0.651203 + 47 | 0.651203 + 48 | 0.651203 + 49 | 0.631477 +(200 rows) + diff --git a/external/smlar/expected/date.out b/external/smlar/expected/date.out new file mode 100644 index 00000000000..880133750c3 --- /dev/null +++ b/external/smlar/expected/date.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + epoch2timestamp(v)::date + FROM + generate_series(1, t) as v + ) AS v + INTO test_date +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::date[]) AS s FROM test_date WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::date[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::date[]) AS s FROM test_date WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::date[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_date ON test_date USING gist (v _date_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::date[]) AS s FROM test_date WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::date[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::date[]) AS s FROM test_date WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::date[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_date; +CREATE INDEX idx_test_date ON test_date USING gin (v _date_sml_ops); +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::date[]) AS s FROM test_date WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::date[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::date[]) AS s FROM test_date WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::date[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/float4.out b/external/smlar/expected/float4.out new file mode 100644 index 00000000000..f44cc0d123f --- /dev/null +++ b/external/smlar/expected/float4.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::float4 + FROM + generate_series(1, t) as v + ) AS v + INTO test_float4 +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_float4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_float4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_float4 ON test_float4 USING gist (v _float4_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_float4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_float4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_float4; +CREATE INDEX idx_test_float4 ON test_float4 USING gin (v _float4_sml_ops); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_float4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_float4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/float8.out b/external/smlar/expected/float8.out new file mode 100644 index 00000000000..bff792c0bea --- /dev/null +++ b/external/smlar/expected/float8.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::float8 + FROM + generate_series(1, t) as v + ) AS v + INTO test_float8 +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_float8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_float8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_float8 ON test_float8 USING gist (v _float8_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_float8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_float8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_float8; +CREATE INDEX idx_test_float8 ON test_float8 USING gin (v _float8_sml_ops); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_float8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_float8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/inet.out b/external/smlar/expected/inet.out new file mode 100644 index 00000000000..b7d47d02c82 --- /dev/null +++ b/external/smlar/expected/inet.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + ( '192.168.1.' || (v % 256) )::inet + FROM + generate_series(1, t) as v + ) AS v + INTO test_inet +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}') AS s FROM test_inet WHERE v % '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}') AS s FROM test_inet WHERE v % '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_inet ON test_inet USING gist (v _inet_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}') AS s FROM test_inet WHERE v % '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}') AS s FROM test_inet WHERE v % '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_inet; +CREATE INDEX idx_test_inet ON test_inet USING gin (v _inet_sml_ops); +SELECT t, smlar(v, '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}') AS s FROM test_inet WHERE v % '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}') AS s FROM test_inet WHERE v % '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/int2.out b/external/smlar/expected/int2.out new file mode 100644 index 00000000000..ed54190fb60 --- /dev/null +++ b/external/smlar/expected/int2.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::int2 + FROM + generate_series(1, t) as v + ) AS v + INTO test_int2 +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int2 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int2 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_int2 ON test_int2 USING gist (v _int2_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int2 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int2 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_int2; +CREATE INDEX idx_test_int2 ON test_int2 USING gin (v _int2_sml_ops); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int2 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int2 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/int4.out b/external/smlar/expected/int4.out new file mode 100644 index 00000000000..90f28ab7965 --- /dev/null +++ b/external/smlar/expected/int4.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::int4 + FROM + generate_series(1, t) as v + ) AS v + INTO test_int4 +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_int4 ON test_int4 USING gist (v _int4_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_int4; +CREATE INDEX idx_test_int4 ON test_int4 USING gin (v _int4_sml_ops); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/int4g.out b/external/smlar/expected/int4g.out new file mode 100644 index 00000000000..433fcd7f47a --- /dev/null +++ b/external/smlar/expected/int4g.out @@ -0,0 +1,1633 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.3); + set_smlar_limit +----------------- + 0.3 +(1 row) + +SET smlar.type = "tfidf"; +SET smlar.stattable = "int4_stat"; +UPDATE test_int4 SET v = v || '{50,50,50}'::int4[] WHERE t >=50 AND t < 60; +INSERT INTO test_int4 VALUES(201, '{50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50}'); +CREATE TABLE int4_stat ( + value int4 UNIQUE, + ndoc int4 NOT NULL CHECK (ndoc>0) +); +INSERT INTO int4_stat ( + SELECT + q.w, + count(*) + FROM + (SELECT array_to_col(v) AS w FROM test_int4 WHERE v IS NOT NULL) AS q + GROUP BY w +); +INSERT INTO int4_stat VALUES (NULL, (SELECT count(*) FROM test_int4)); +DROP INDEX idx_test_int4; +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306753 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +---------- + 0.997241 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.398379 + 51 | 0.394635 + 52 | 0.390962 + 53 | 0.387358 + 54 | 0.38382 + 55 | 0.380346 + 56 | 0.376933 + 57 | 0.373581 + 58 | 0.370286 + 60 | 0.369012 + 59 | 0.367048 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 50 | 0.510326 + 51 | 0.50553 + 52 | 0.500825 + 18 | 0.496591 + 53 | 0.496208 + 54 | 0.491676 + 55 | 0.487225 + 56 | 0.482854 + 19 | 0.482412 + 57 | 0.47856 + 58 | 0.47434 + 59 | 0.470191 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306662 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +---------- + 0.999075 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 50 | 0.419425 + 49 | 0.417358 + 51 | 0.41482 + 52 | 0.410324 + 53 | 0.405932 + 54 | 0.40164 + 55 | 0.397443 + 56 | 0.393339 + 57 | 0.389323 + 58 | 0.385391 + 59 | 0.381541 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 50 | 0.481241 + 51 | 0.475957 + 52 | 0.470799 + 20 | 0.469285 + 53 | 0.465759 + 54 | 0.460835 + 21 | 0.457083 + 55 | 0.45602 + 4 | 0.45545 + 56 | 0.45131 + 57 | 0.446702 + 22 | 0.4457 + 58 | 0.442191 + 59 | 0.437774 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306545 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.413885 + 51 | 0.408856 + 52 | 0.403962 + 53 | 0.399199 + 54 | 0.394558 + 55 | 0.390035 + 56 | 0.385625 + 57 | 0.381322 + 58 | 0.377121 + 59 | 0.373019 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 50 | 0.425793 + 24 | 0.425047 + 51 | 0.420619 + 33 | 0.419273 + 25 | 0.415635 + 52 | 0.415585 + 34 | 0.41221 + 53 | 0.410684 + 26 | 0.406753 + 54 | 0.40591 + 35 | 0.405439 + 55 | 0.401257 + 36 | 0.398938 + 27 | 0.398352 + 56 | 0.39672 + 37 | 0.392689 + 57 | 0.392293 + 28 | 0.390389 + 49 | 0.390116 + 58 | 0.387972 + 38 | 0.386675 + 59 | 0.383752 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +CREATE INDEX idx_test_int4 ON test_int4 USING gist (v _int4_sml_ops); +SET enable_seqscan = off; +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306753 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +---------- + 0.997241 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.398379 + 51 | 0.394635 + 52 | 0.390962 + 53 | 0.387358 + 54 | 0.38382 + 55 | 0.380346 + 56 | 0.376933 + 57 | 0.373581 + 58 | 0.370286 + 60 | 0.369012 + 59 | 0.367048 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 50 | 0.510326 + 51 | 0.50553 + 52 | 0.500825 + 18 | 0.496591 + 53 | 0.496208 + 54 | 0.491676 + 55 | 0.487225 + 56 | 0.482854 + 19 | 0.482412 + 57 | 0.47856 + 58 | 0.47434 + 59 | 0.470191 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306662 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +---------- + 0.999075 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 50 | 0.419425 + 49 | 0.417358 + 51 | 0.41482 + 52 | 0.410324 + 53 | 0.405932 + 54 | 0.40164 + 55 | 0.397443 + 56 | 0.393339 + 57 | 0.389323 + 58 | 0.385391 + 59 | 0.381541 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 50 | 0.481241 + 51 | 0.475957 + 52 | 0.470799 + 20 | 0.469285 + 53 | 0.465759 + 54 | 0.460835 + 21 | 0.457083 + 55 | 0.45602 + 4 | 0.45545 + 56 | 0.45131 + 57 | 0.446702 + 22 | 0.4457 + 58 | 0.442191 + 59 | 0.437774 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306545 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.413885 + 51 | 0.408856 + 52 | 0.403962 + 53 | 0.399199 + 54 | 0.394558 + 55 | 0.390035 + 56 | 0.385625 + 57 | 0.381322 + 58 | 0.377121 + 59 | 0.373019 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 50 | 0.425793 + 24 | 0.425047 + 51 | 0.420619 + 33 | 0.419273 + 25 | 0.415635 + 52 | 0.415585 + 34 | 0.41221 + 53 | 0.410684 + 26 | 0.406753 + 54 | 0.40591 + 35 | 0.405439 + 55 | 0.401257 + 36 | 0.398938 + 27 | 0.398352 + 56 | 0.39672 + 37 | 0.392689 + 57 | 0.392293 + 28 | 0.390389 + 49 | 0.390116 + 58 | 0.387972 + 38 | 0.386675 + 59 | 0.383752 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +DROP INDEX idx_test_int4; +CREATE INDEX idx_test_int4 ON test_int4 USING gin (v _int4_sml_ops); +SET enable_seqscan = off; +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +---------- + 0.997241 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +---------- + 0.999075 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306545 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.413885 + 51 | 0.408856 + 52 | 0.403962 + 53 | 0.399199 + 54 | 0.394558 + 55 | 0.390035 + 56 | 0.385625 + 57 | 0.381322 + 58 | 0.377121 + 59 | 0.373019 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 50 | 0.425793 + 24 | 0.425047 + 51 | 0.420619 + 33 | 0.419273 + 25 | 0.415635 + 52 | 0.415585 + 34 | 0.41221 + 53 | 0.410684 + 26 | 0.406753 + 54 | 0.40591 + 35 | 0.405439 + 55 | 0.401257 + 36 | 0.398938 + 27 | 0.398352 + 56 | 0.39672 + 37 | 0.392689 + 57 | 0.392293 + 28 | 0.390389 + 49 | 0.390116 + 58 | 0.387972 + 38 | 0.386675 + 59 | 0.383752 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +RESET enable_seqscan; diff --git a/external/smlar/expected/int4i.out b/external/smlar/expected/int4i.out new file mode 100644 index 00000000000..92a82461d7e --- /dev/null +++ b/external/smlar/expected/int4i.out @@ -0,0 +1,42 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(5.0); + set_smlar_limit +----------------- + 5 +(1 row) + +SET smlar.type='overlap'; +DROP INDEX idx_test_int4; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); + smlar +------- + 2 +(1 row) + +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_int4 WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + t | s +-----+---- + 199 | 10 + 200 | 10 + 198 | 9 + 197 | 8 + 196 | 7 + 195 | 6 + 194 | 5 +(7 rows) + +CREATE INDEX idx_test_int4 ON test_int4 USING gin (v _int4_sml_ops); +SET enable_seqscan = off; +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_int4 WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + t | s +-----+---- + 199 | 10 + 200 | 10 + 198 | 9 + 197 | 8 + 196 | 7 + 195 | 6 + 194 | 5 +(7 rows) + +RESET enable_seqscan; diff --git a/external/smlar/expected/int8.out b/external/smlar/expected/int8.out new file mode 100644 index 00000000000..7fdaebe4d6c --- /dev/null +++ b/external/smlar/expected/int8.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::int8 + FROM + generate_series(1, t) as v + ) AS v + INTO test_int8 +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_int8 ON test_int8 USING gist (v _int8_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_int8; +CREATE INDEX idx_test_int8 ON test_int8 USING gin (v _int8_sml_ops); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/int8g.out b/external/smlar/expected/int8g.out new file mode 100644 index 00000000000..b38ba2cd6c2 --- /dev/null +++ b/external/smlar/expected/int8g.out @@ -0,0 +1,1633 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.3); + set_smlar_limit +----------------- + 0.3 +(1 row) + +SET smlar.type = "tfidf"; +SET smlar.stattable = "int8_stat"; +UPDATE test_int8 SET v = v || '{50,50,50}'::int8[] WHERE t >=50 AND t < 60; +INSERT INTO test_int8 VALUES(201, '{50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50}'); +CREATE TABLE int8_stat ( + value int8 UNIQUE, + ndoc int4 NOT NULL CHECK (ndoc>0) +); +INSERT INTO int8_stat ( + SELECT + q.w, + count(*) + FROM + (SELECT array_to_col(v) AS w FROM test_int8 WHERE v IS NOT NULL) AS q + GROUP BY w +); +INSERT INTO int8_stat VALUES (NULL, (SELECT count(*) FROM test_int8)); +DROP INDEX idx_test_int8; +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306753 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +---------- + 0.997241 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.398379 + 51 | 0.394635 + 52 | 0.390962 + 53 | 0.387358 + 54 | 0.38382 + 55 | 0.380346 + 56 | 0.376933 + 57 | 0.373581 + 58 | 0.370286 + 60 | 0.369012 + 59 | 0.367048 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 50 | 0.510326 + 51 | 0.50553 + 52 | 0.500825 + 18 | 0.496591 + 53 | 0.496208 + 54 | 0.491676 + 55 | 0.487225 + 56 | 0.482854 + 19 | 0.482412 + 57 | 0.47856 + 58 | 0.47434 + 59 | 0.470191 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306662 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +---------- + 0.999075 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 50 | 0.419425 + 49 | 0.417358 + 51 | 0.41482 + 52 | 0.410324 + 53 | 0.405932 + 54 | 0.40164 + 55 | 0.397443 + 56 | 0.393339 + 57 | 0.389323 + 58 | 0.385391 + 59 | 0.381541 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 50 | 0.481241 + 51 | 0.475957 + 52 | 0.470799 + 20 | 0.469285 + 53 | 0.465759 + 54 | 0.460835 + 21 | 0.457083 + 55 | 0.45602 + 4 | 0.45545 + 56 | 0.45131 + 57 | 0.446702 + 22 | 0.4457 + 58 | 0.442191 + 59 | 0.437774 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306545 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.413885 + 51 | 0.408856 + 52 | 0.403962 + 53 | 0.399199 + 54 | 0.394558 + 55 | 0.390035 + 56 | 0.385625 + 57 | 0.381322 + 58 | 0.377121 + 59 | 0.373019 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 50 | 0.425793 + 24 | 0.425047 + 51 | 0.420619 + 33 | 0.419273 + 25 | 0.415635 + 52 | 0.415585 + 34 | 0.41221 + 53 | 0.410684 + 26 | 0.406753 + 54 | 0.40591 + 35 | 0.405439 + 55 | 0.401257 + 36 | 0.398938 + 27 | 0.398352 + 56 | 0.39672 + 37 | 0.392689 + 57 | 0.392293 + 28 | 0.390389 + 49 | 0.390116 + 58 | 0.387972 + 38 | 0.386675 + 59 | 0.383752 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +CREATE INDEX idx_test_int8 ON test_int8 USING gist (v _int8_sml_ops); +SET enable_seqscan = off; +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306753 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +---------- + 0.997241 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.398379 + 51 | 0.394635 + 52 | 0.390962 + 53 | 0.387358 + 54 | 0.38382 + 55 | 0.380346 + 56 | 0.376933 + 57 | 0.373581 + 58 | 0.370286 + 60 | 0.369012 + 59 | 0.367048 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 50 | 0.510326 + 51 | 0.50553 + 52 | 0.500825 + 18 | 0.496591 + 53 | 0.496208 + 54 | 0.491676 + 55 | 0.487225 + 56 | 0.482854 + 19 | 0.482412 + 57 | 0.47856 + 58 | 0.47434 + 59 | 0.470191 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306662 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +---------- + 0.999075 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 50 | 0.419425 + 49 | 0.417358 + 51 | 0.41482 + 52 | 0.410324 + 53 | 0.405932 + 54 | 0.40164 + 55 | 0.397443 + 56 | 0.393339 + 57 | 0.389323 + 58 | 0.385391 + 59 | 0.381541 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 50 | 0.481241 + 51 | 0.475957 + 52 | 0.470799 + 20 | 0.469285 + 53 | 0.465759 + 54 | 0.460835 + 21 | 0.457083 + 55 | 0.45602 + 4 | 0.45545 + 56 | 0.45131 + 57 | 0.446702 + 22 | 0.4457 + 58 | 0.442191 + 59 | 0.437774 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306545 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.413885 + 51 | 0.408856 + 52 | 0.403962 + 53 | 0.399199 + 54 | 0.394558 + 55 | 0.390035 + 56 | 0.385625 + 57 | 0.381322 + 58 | 0.377121 + 59 | 0.373019 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 50 | 0.425793 + 24 | 0.425047 + 51 | 0.420619 + 33 | 0.419273 + 25 | 0.415635 + 52 | 0.415585 + 34 | 0.41221 + 53 | 0.410684 + 26 | 0.406753 + 54 | 0.40591 + 35 | 0.405439 + 55 | 0.401257 + 36 | 0.398938 + 27 | 0.398352 + 56 | 0.39672 + 37 | 0.392689 + 57 | 0.392293 + 28 | 0.390389 + 49 | 0.390116 + 58 | 0.387972 + 38 | 0.386675 + 59 | 0.383752 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +DROP INDEX idx_test_int8; +CREATE INDEX idx_test_int8 ON test_int8 USING gin (v _int8_sml_ops); +SET enable_seqscan = off; +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +---------- + 0.997241 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +---------- + 0.999075 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306545 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.413885 + 51 | 0.408856 + 52 | 0.403962 + 53 | 0.399199 + 54 | 0.394558 + 55 | 0.390035 + 56 | 0.385625 + 57 | 0.381322 + 58 | 0.377121 + 59 | 0.373019 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 50 | 0.425793 + 24 | 0.425047 + 51 | 0.420619 + 33 | 0.419273 + 25 | 0.415635 + 52 | 0.415585 + 34 | 0.41221 + 53 | 0.410684 + 26 | 0.406753 + 54 | 0.40591 + 35 | 0.405439 + 55 | 0.401257 + 36 | 0.398938 + 27 | 0.398352 + 56 | 0.39672 + 37 | 0.392689 + 57 | 0.392293 + 28 | 0.390389 + 49 | 0.390116 + 58 | 0.387972 + 38 | 0.386675 + 59 | 0.383752 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +RESET enable_seqscan; diff --git a/external/smlar/expected/int8i.out b/external/smlar/expected/int8i.out new file mode 100644 index 00000000000..201cf63a14a --- /dev/null +++ b/external/smlar/expected/int8i.out @@ -0,0 +1,42 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(5.0); + set_smlar_limit +----------------- + 5 +(1 row) + +SET smlar.type='overlap'; +DROP INDEX idx_test_int8; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); + smlar +------- + 2 +(1 row) + +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_int8 WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + t | s +-----+---- + 199 | 10 + 200 | 10 + 198 | 9 + 197 | 8 + 196 | 7 + 195 | 6 + 194 | 5 +(7 rows) + +CREATE INDEX idx_test_int8 ON test_int8 USING gin (v _int8_sml_ops); +SET enable_seqscan = off; +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_int8 WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + t | s +-----+---- + 199 | 10 + 200 | 10 + 198 | 9 + 197 | 8 + 196 | 7 + 195 | 6 + 194 | 5 +(7 rows) + +RESET enable_seqscan; diff --git a/external/smlar/expected/interval.out b/external/smlar/expected/interval.out new file mode 100644 index 00000000000..1d33b6ea548 --- /dev/null +++ b/external/smlar/expected/interval.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::text::interval + FROM + generate_series(1, t) as v + ) AS v + INTO test_interval +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_interval ON test_interval USING gist (v _interval_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_interval; +CREATE INDEX idx_test_interval ON test_interval USING gin (v _interval_sml_ops); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/intervalg.out b/external/smlar/expected/intervalg.out new file mode 100644 index 00000000000..7122a900dc4 --- /dev/null +++ b/external/smlar/expected/intervalg.out @@ -0,0 +1,1633 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.3); + set_smlar_limit +----------------- + 0.3 +(1 row) + +SET smlar.type = "tfidf"; +SET smlar.stattable = "interval_stat"; +UPDATE test_interval SET v = v || '{50,50,50}'::interval[] WHERE t >=50 AND t < 60; +INSERT INTO test_interval VALUES(201, '{50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50}'); +CREATE TABLE interval_stat ( + value interval UNIQUE, + ndoc int4 NOT NULL CHECK (ndoc>0) +); +INSERT INTO interval_stat ( + SELECT + q.w, + count(*) + FROM + (SELECT array_to_col(v) AS w FROM test_interval WHERE v IS NOT NULL) AS q + GROUP BY w +); +INSERT INTO interval_stat VALUES (NULL, (SELECT count(*) FROM test_interval)); +DROP INDEX idx_test_interval; +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306753 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +---------- + 0.997241 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.398379 + 51 | 0.394635 + 52 | 0.390962 + 53 | 0.387358 + 54 | 0.38382 + 55 | 0.380346 + 56 | 0.376933 + 57 | 0.373581 + 58 | 0.370286 + 60 | 0.369012 + 59 | 0.367048 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 50 | 0.510326 + 51 | 0.50553 + 52 | 0.500825 + 18 | 0.496591 + 53 | 0.496208 + 54 | 0.491676 + 55 | 0.487225 + 56 | 0.482854 + 19 | 0.482412 + 57 | 0.47856 + 58 | 0.47434 + 59 | 0.470191 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306662 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +---------- + 0.999075 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 50 | 0.419425 + 49 | 0.417358 + 51 | 0.41482 + 52 | 0.410324 + 53 | 0.405932 + 54 | 0.40164 + 55 | 0.397443 + 56 | 0.393339 + 57 | 0.389323 + 58 | 0.385391 + 59 | 0.381541 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 50 | 0.481241 + 51 | 0.475957 + 52 | 0.470799 + 20 | 0.469285 + 53 | 0.465759 + 54 | 0.460835 + 21 | 0.457083 + 55 | 0.45602 + 4 | 0.45545 + 56 | 0.45131 + 57 | 0.446702 + 22 | 0.4457 + 58 | 0.442191 + 59 | 0.437774 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306545 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.413885 + 51 | 0.408856 + 52 | 0.403962 + 53 | 0.399199 + 54 | 0.394558 + 55 | 0.390035 + 56 | 0.385625 + 57 | 0.381322 + 58 | 0.377121 + 59 | 0.373019 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 50 | 0.425793 + 24 | 0.425047 + 51 | 0.420619 + 33 | 0.419273 + 25 | 0.415635 + 52 | 0.415585 + 34 | 0.41221 + 53 | 0.410684 + 26 | 0.406753 + 54 | 0.40591 + 35 | 0.405439 + 55 | 0.401257 + 36 | 0.398938 + 27 | 0.398352 + 56 | 0.39672 + 37 | 0.392689 + 57 | 0.392293 + 28 | 0.390389 + 49 | 0.390116 + 58 | 0.387972 + 38 | 0.386675 + 59 | 0.383752 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +CREATE INDEX idx_test_interval ON test_interval USING gist (v _interval_sml_ops); +SET enable_seqscan = off; +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306753 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +---------- + 0.997241 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.398379 + 51 | 0.394635 + 52 | 0.390962 + 53 | 0.387358 + 54 | 0.38382 + 55 | 0.380346 + 56 | 0.376933 + 57 | 0.373581 + 58 | 0.370286 + 60 | 0.369012 + 59 | 0.367048 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 50 | 0.510326 + 51 | 0.50553 + 52 | 0.500825 + 18 | 0.496591 + 53 | 0.496208 + 54 | 0.491676 + 55 | 0.487225 + 56 | 0.482854 + 19 | 0.482412 + 57 | 0.47856 + 58 | 0.47434 + 59 | 0.470191 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306662 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +---------- + 0.999075 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 50 | 0.419425 + 49 | 0.417358 + 51 | 0.41482 + 52 | 0.410324 + 53 | 0.405932 + 54 | 0.40164 + 55 | 0.397443 + 56 | 0.393339 + 57 | 0.389323 + 58 | 0.385391 + 59 | 0.381541 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 50 | 0.481241 + 51 | 0.475957 + 52 | 0.470799 + 20 | 0.469285 + 53 | 0.465759 + 54 | 0.460835 + 21 | 0.457083 + 55 | 0.45602 + 4 | 0.45545 + 56 | 0.45131 + 57 | 0.446702 + 22 | 0.4457 + 58 | 0.442191 + 59 | 0.437774 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306545 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.413885 + 51 | 0.408856 + 52 | 0.403962 + 53 | 0.399199 + 54 | 0.394558 + 55 | 0.390035 + 56 | 0.385625 + 57 | 0.381322 + 58 | 0.377121 + 59 | 0.373019 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 50 | 0.425793 + 24 | 0.425047 + 51 | 0.420619 + 33 | 0.419273 + 25 | 0.415635 + 52 | 0.415585 + 34 | 0.41221 + 53 | 0.410684 + 26 | 0.406753 + 54 | 0.40591 + 35 | 0.405439 + 55 | 0.401257 + 36 | 0.398938 + 27 | 0.398352 + 56 | 0.39672 + 37 | 0.392689 + 57 | 0.392293 + 28 | 0.390389 + 49 | 0.390116 + 58 | 0.387972 + 38 | 0.386675 + 59 | 0.383752 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +DROP INDEX idx_test_interval; +CREATE INDEX idx_test_interval ON test_interval USING gin (v _interval_sml_ops); +SET enable_seqscan = off; +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +---------- + 0.997241 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +---------- + 0.999075 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306545 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.413885 + 51 | 0.408856 + 52 | 0.403962 + 53 | 0.399199 + 54 | 0.394558 + 55 | 0.390035 + 56 | 0.385625 + 57 | 0.381322 + 58 | 0.377121 + 59 | 0.373019 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 50 | 0.425793 + 24 | 0.425047 + 51 | 0.420619 + 33 | 0.419273 + 25 | 0.415635 + 52 | 0.415585 + 34 | 0.41221 + 53 | 0.410684 + 26 | 0.406753 + 54 | 0.40591 + 35 | 0.405439 + 55 | 0.401257 + 36 | 0.398938 + 27 | 0.398352 + 56 | 0.39672 + 37 | 0.392689 + 57 | 0.392293 + 28 | 0.390389 + 49 | 0.390116 + 58 | 0.387972 + 38 | 0.386675 + 59 | 0.383752 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +RESET enable_seqscan; diff --git a/external/smlar/expected/intervali.out b/external/smlar/expected/intervali.out new file mode 100644 index 00000000000..fd1f5b8467c --- /dev/null +++ b/external/smlar/expected/intervali.out @@ -0,0 +1,42 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(5.0); + set_smlar_limit +----------------- + 5 +(1 row) + +SET smlar.type='overlap'; +DROP INDEX idx_test_interval; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); + smlar +------- + 2 +(1 row) + +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_interval WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + t | s +-----+---- + 199 | 10 + 200 | 10 + 198 | 9 + 197 | 8 + 196 | 7 + 195 | 6 + 194 | 5 +(7 rows) + +CREATE INDEX idx_test_interval ON test_interval USING gin (v _interval_sml_ops); +SET enable_seqscan = off; +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_interval WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + t | s +-----+---- + 199 | 10 + 200 | 10 + 198 | 9 + 197 | 8 + 196 | 7 + 195 | 6 + 194 | 5 +(7 rows) + +RESET enable_seqscan; diff --git a/external/smlar/expected/macaddr.out b/external/smlar/expected/macaddr.out new file mode 100644 index 00000000000..14cac5f56aa --- /dev/null +++ b/external/smlar/expected/macaddr.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + ('01:01:01:01:01:' || (v % 100))::macaddr + FROM + generate_series(1, t) as v + ) AS v + INTO test_macaddr +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{01:01:01:01:01:10,01:01:01:01:01:9,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:3,01:01:01:01:01:2,01:01:01:01:01:1}') AS s FROM test_macaddr WHERE v % '{01:01:01:01:01:10,01:01:01:01:01:9,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:3,01:01:01:01:01:2,01:01:01:01:01:1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{01:01:01:01:01:50,01:01:01:01:01:49,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:33,01:01:01:01:01:2,01:01:01:01:01:1}') AS s FROM test_macaddr WHERE v % '{01:01:01:01:01:50,01:01:01:01:01:49,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:33,01:01:01:01:01:2,01:01:01:01:01:1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_macaddr ON test_macaddr USING gist (v _macaddr_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{01:01:01:01:01:10,01:01:01:01:01:9,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:3,01:01:01:01:01:2,01:01:01:01:01:1}') AS s FROM test_macaddr WHERE v % '{01:01:01:01:01:10,01:01:01:01:01:9,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:3,01:01:01:01:01:2,01:01:01:01:01:1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{01:01:01:01:01:50,01:01:01:01:01:49,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:33,01:01:01:01:01:2,01:01:01:01:01:1}') AS s FROM test_macaddr WHERE v % '{01:01:01:01:01:50,01:01:01:01:01:49,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:33,01:01:01:01:01:2,01:01:01:01:01:1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_macaddr; +CREATE INDEX idx_test_macaddr ON test_macaddr USING gin (v _macaddr_sml_ops); +SELECT t, smlar(v, '{01:01:01:01:01:10,01:01:01:01:01:9,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:3,01:01:01:01:01:2,01:01:01:01:01:1}') AS s FROM test_macaddr WHERE v % '{01:01:01:01:01:10,01:01:01:01:01:9,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:3,01:01:01:01:01:2,01:01:01:01:01:1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{01:01:01:01:01:50,01:01:01:01:01:49,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:33,01:01:01:01:01:2,01:01:01:01:01:1}') AS s FROM test_macaddr WHERE v % '{01:01:01:01:01:50,01:01:01:01:01:49,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:33,01:01:01:01:01:2,01:01:01:01:01:1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/money.out b/external/smlar/expected/money.out new file mode 100644 index 00000000000..d3f4ae1b763 --- /dev/null +++ b/external/smlar/expected/money.out @@ -0,0 +1,105 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::text::money + FROM + generate_series(1, t) as v + ) AS v + INTO test_money +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_money WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_money WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_money ON test_money USING gin (v _money_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_money WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_money WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/numeric.out b/external/smlar/expected/numeric.out new file mode 100644 index 00000000000..cf8882fa7e3 --- /dev/null +++ b/external/smlar/expected/numeric.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::numeric + FROM + generate_series(1, t) as v + ) AS v + INTO test_numeric +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_numeric WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_numeric WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_numeric ON test_numeric USING gist (v _numeric_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_numeric WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_numeric WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_numeric; +CREATE INDEX idx_test_numeric ON test_numeric USING gin (v _numeric_sml_ops); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_numeric WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_numeric WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/oid.out b/external/smlar/expected/oid.out new file mode 100644 index 00000000000..ac67cf79b79 --- /dev/null +++ b/external/smlar/expected/oid.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::oid + FROM + generate_series(1, t) as v + ) AS v + INTO test_oid +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_oid WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_oid WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_oid ON test_oid USING gist (v _oid_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_oid WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_oid WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_oid; +CREATE INDEX idx_test_oid ON test_oid USING gin (v _oid_sml_ops); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_oid WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_oid WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/smlar.out b/external/smlar/expected/smlar.out new file mode 100644 index 00000000000..5dac034166b --- /dev/null +++ b/external/smlar/expected/smlar.out @@ -0,0 +1,355 @@ +CREATE EXTENSION smlar; +set extra_float_digits =0; +--sanity check +SELECT + opc.opcname, + t.typname, + opc.opcdefault +FROM + pg_opclass opc, + pg_am am, + pg_type t, + pg_type k +WHERE + opc.opcmethod = am.oid AND + am.amname='gist' AND + opc.opcintype = t.oid AND + opc.opckeytype = k.oid AND + k.typname='gsmlsign' +ORDER BY opc.opcname; + opcname | typname | opcdefault +----------------------+--------------+------------ + _bytea_sml_ops | _bytea | f + _char_sml_ops | _char | f + _cidr_sml_ops | _cidr | f + _date_sml_ops | _date | f + _float4_sml_ops | _float4 | f + _float8_sml_ops | _float8 | f + _inet_sml_ops | _inet | f + _int2_sml_ops | _int2 | f + _int4_sml_ops | _int4 | f + _int8_sml_ops | _int8 | f + _interval_sml_ops | _interval | f + _macaddr_sml_ops | _macaddr | f + _numeric_sml_ops | _numeric | f + _oid_sml_ops | _oid | f + _text_sml_ops | _text | f + _time_sml_ops | _time | f + _timestamp_sml_ops | _timestamp | f + _timestamptz_sml_ops | _timestamptz | f + _timetz_sml_ops | _timetz | f + _varchar_sml_ops | _varchar | f +(20 rows) + +SELECT + opc.opcname, + t.typname, + opc.opcdefault +FROM + pg_opclass opc, + pg_am am, + pg_type t +WHERE + opc.opcmethod = am.oid AND + am.amname='gin' AND + opc.opcintype = t.oid AND + opc.opcname ~ '_sml_ops$' +ORDER BY opc.opcname; + opcname | typname | opcdefault +----------------------+--------------+------------ + _bit_sml_ops | _bit | f + _bytea_sml_ops | _bytea | f + _char_sml_ops | _char | f + _cidr_sml_ops | _cidr | f + _date_sml_ops | _date | f + _float4_sml_ops | _float4 | f + _float8_sml_ops | _float8 | f + _inet_sml_ops | _inet | f + _int2_sml_ops | _int2 | f + _int4_sml_ops | _int4 | f + _int8_sml_ops | _int8 | f + _interval_sml_ops | _interval | f + _macaddr_sml_ops | _macaddr | f + _money_sml_ops | _money | f + _numeric_sml_ops | _numeric | f + _oid_sml_ops | _oid | f + _text_sml_ops | _text | f + _time_sml_ops | _time | f + _timestamp_sml_ops | _timestamp | f + _timestamptz_sml_ops | _timestamptz | f + _timetz_sml_ops | _timetz | f + _varbit_sml_ops | _varbit | f + _varchar_sml_ops | _varchar | f +(23 rows) + +SELECT + trim( leading '_' from t.typname ) || '[]' AS "Array Type", + gin.opcname AS "GIN operator class", + gist.opcname AS "GiST operator class" +FROM + ( + SELECT * + FROM + pg_catalog.pg_opclass, + pg_catalog.pg_am + WHERE + pg_opclass.opcmethod = pg_am.oid AND + pg_am.amname = 'gin' AND + pg_opclass.opcname ~ '_sml_ops$' + ) AS gin + FULL OUTER JOIN + ( + SELECT * + FROM + pg_catalog.pg_opclass, + pg_catalog.pg_am + WHERE + pg_opclass.opcmethod = pg_am.oid AND + pg_am.amname = 'gist' AND + pg_opclass.opcname ~ '_sml_ops$' + ) AS gist + ON ( + gist.opcname = gin.opcname AND + gist.opcintype = gin.opcintype + ), + pg_catalog.pg_type t +WHERE + t.oid = COALESCE(gist.opcintype, gin.opcintype) AND + t.typarray = 0 +ORDER BY + "Array Type" ASC +; + Array Type | GIN operator class | GiST operator class +---------------+----------------------+---------------------- + bit[] | _bit_sml_ops | + bytea[] | _bytea_sml_ops | _bytea_sml_ops + char[] | _char_sml_ops | _char_sml_ops + cidr[] | _cidr_sml_ops | _cidr_sml_ops + date[] | _date_sml_ops | _date_sml_ops + float4[] | _float4_sml_ops | _float4_sml_ops + float8[] | _float8_sml_ops | _float8_sml_ops + inet[] | _inet_sml_ops | _inet_sml_ops + int2[] | _int2_sml_ops | _int2_sml_ops + int4[] | _int4_sml_ops | _int4_sml_ops + int8[] | _int8_sml_ops | _int8_sml_ops + interval[] | _interval_sml_ops | _interval_sml_ops + macaddr[] | _macaddr_sml_ops | _macaddr_sml_ops + money[] | _money_sml_ops | + numeric[] | _numeric_sml_ops | _numeric_sml_ops + oid[] | _oid_sml_ops | _oid_sml_ops + text[] | _text_sml_ops | _text_sml_ops + time[] | _time_sml_ops | _time_sml_ops + timestamp[] | _timestamp_sml_ops | _timestamp_sml_ops + timestamptz[] | _timestamptz_sml_ops | _timestamptz_sml_ops + timetz[] | _timetz_sml_ops | _timetz_sml_ops + varbit[] | _varbit_sml_ops | + varchar[] | _varchar_sml_ops | _varchar_sml_ops +(23 rows) + +SELECT set_smlar_limit(0.1); + set_smlar_limit +----------------- + 0.1 +(1 row) + +SET smlar.threshold = 0.6; +--First checks +SELECT smlar('{3,2}'::int[], '{3,2,1}'); + smlar +---------- + 0.816497 +(1 row) + +SELECT smlar('{2,1}'::int[], '{3,2,1}'); + smlar +---------- + 0.816497 +(1 row) + +SELECT smlar('{}'::int[], '{3,2,1}'); + smlar +------- + 0 +(1 row) + +SELECT smlar('{12,10}'::int[], '{3,2,1}'); + smlar +------- + 0 +(1 row) + +SELECT smlar('{123}'::int[], '{}'); + smlar +------- + 0 +(1 row) + +SELECT smlar('{1,4,6}'::int[], '{1,4,6}'); + smlar +------- + 1 +(1 row) + +SELECT smlar('{1,4,6}'::int[], '{6,4,1}'); + smlar +------- + 1 +(1 row) + +SELECT smlar('{1,4,6}'::int[], '{5,4,6}'); + smlar +---------- + 0.666667 +(1 row) + +SELECT smlar('{1,4,6}'::int[], '{5,4,6}'); + smlar +---------- + 0.666667 +(1 row) + +SELECT smlar('{1,2}'::int[], '{2,2,2,2,2,1}'); + smlar +------- + 1 +(1 row) + +SELECT smlar('{1,2}'::int[], '{1,2,2,2,2,2}'); + smlar +------- + 1 +(1 row) + +SELECT smlar('{}'::int[], '{3}'); + smlar +------- + 0 +(1 row) + +SELECT smlar('{3}'::int[], '{3}'); + smlar +------- + 1 +(1 row) + +SELECT smlar('{2}'::int[], '{3}'); + smlar +------- + 0 +(1 row) + +SELECT smlar('{1,4,6}'::int[], '{5,4,6}', 'N.i / (N.a + N.b)' ); + smlar +---------- + 0.333333 +(1 row) + +SELECT smlar('{1,4,6}'::int[], '{5,4,6}', 'N.i / sqrt(N.a * N.b)' ); + smlar +---------- + 0.666667 +(1 row) + +SELECT tsvector2textarray('qwe:12 asd:45'); + tsvector2textarray +-------------------- + {asd,qwe} +(1 row) + +SELECT array_unique('{12,12,1,4,1,16}'::int4[]); + array_unique +-------------- + {1,4,12,16} +(1 row) + +SELECT array_unique('{12,12,1,4,1,16}'::bigint[]); + array_unique +-------------- + {1,4,12,16} +(1 row) + +SELECT array_unique('{12,12,1,4,1,16}'::smallint[]); + array_unique +-------------- + {1,4,12,16} +(1 row) + +SELECT array_unique('{12,12,1,4,1,16}'::text[]); + array_unique +-------------- + {1,12,16,4} +(1 row) + +SELECT array_unique('{12,12,1,4,1,16}'::varchar[]); + array_unique +-------------- + {1,12,16,4} +(1 row) + +SELECT inarray('{12,12,1,4,1,16}'::int[], 13::int); + inarray +--------- + 0 +(1 row) + +SELECT inarray('{12,12,1,4,1,16}'::int[], 12::int); + inarray +--------- + 1 +(1 row) + +SELECT inarray('{12,12,1,4,1,16}'::text[], 13::text); + inarray +--------- + 0 +(1 row) + +SELECT inarray('{12,12,1,4,1,16}'::text[], 12::text); + inarray +--------- + 1 +(1 row) + +SELECT inarray('{12,12,1,4,1,16}'::int[], 13::int, 0.9, 0.1); + inarray +--------- + 0.1 +(1 row) + +SELECT inarray('{12,12,1,4,1,16}'::int[], 12::int, 0.9, 0.1); + inarray +--------- + 0.9 +(1 row) + +SELECT inarray('{12,12,1,4,1,16}'::text[], 13::text, 0.9, 0.1); + inarray +--------- + 0.1 +(1 row) + +SELECT inarray('{12,12,1,4,1,16}'::text[], 12::text, 0.9, 0.1); + inarray +--------- + 0.9 +(1 row) + +--testing function +CREATE OR REPLACE FUNCTION epoch2timestamp(int4) +RETURNS timestamp AS $$ + SELECT ('1970-01-01 00:00:00'::timestamp + ( ($1*3600*24 + $1) ::text || ' seconds' )::interval)::timestamp; + $$ LANGUAGE SQL RETURNS NULL ON NULL INPUT IMMUTABLE; +CREATE OR REPLACE FUNCTION to_tsp_array(_int4) +RETURNS _timestamp AS $$ + SELECT ARRAY( + SELECT + epoch2timestamp( $1[n] ) + FROM + generate_series( 1, array_upper( $1, 1) - array_lower( $1, 1 ) + 1 ) AS n + ); + $$ LANGUAGE SQL RETURNS NULL ON NULL INPUT IMMUTABLE; +CREATE OR REPLACE FUNCTION array_to_col(anyarray) +RETURNS SETOF anyelement AS +$$ + SELECT $1[n] FROM generate_series( 1, array_upper( $1, 1) - array_lower( $1, 1 ) + 1 ) AS n; +$$ LANGUAGE SQL RETURNS NULL ON NULL INPUT IMMUTABLE; diff --git a/external/smlar/expected/text.out b/external/smlar/expected/text.out new file mode 100644 index 00000000000..56cbd36d030 --- /dev/null +++ b/external/smlar/expected/text.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::text + FROM + generate_series(1, t) as v + ) AS v + INTO test_text +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_text ON test_text USING gist (v _text_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_text; +CREATE INDEX idx_test_text ON test_text USING gin (v _text_sml_ops); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/textg.out b/external/smlar/expected/textg.out new file mode 100644 index 00000000000..38fc5781e8e --- /dev/null +++ b/external/smlar/expected/textg.out @@ -0,0 +1,1633 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.3); + set_smlar_limit +----------------- + 0.3 +(1 row) + +SET smlar.type = "tfidf"; +SET smlar.stattable = "text_stat"; +UPDATE test_text SET v = v || '{50,50,50}'::text[] WHERE t >=50 AND t < 60; +INSERT INTO test_text VALUES(201, '{50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50}'); +CREATE TABLE text_stat ( + value text UNIQUE, + ndoc int4 NOT NULL CHECK (ndoc>0) +); +INSERT INTO text_stat ( + SELECT + q.w, + count(*) + FROM + (SELECT array_to_col(v) AS w FROM test_text WHERE v IS NOT NULL) AS q + GROUP BY w +); +INSERT INTO text_stat VALUES (NULL, (SELECT count(*) FROM test_text)); +DROP INDEX idx_test_text; +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306753 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +---------- + 0.997241 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.398379 + 51 | 0.394635 + 52 | 0.390962 + 53 | 0.387358 + 54 | 0.38382 + 55 | 0.380346 + 56 | 0.376933 + 57 | 0.373581 + 58 | 0.370286 + 60 | 0.369012 + 59 | 0.367048 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 50 | 0.510326 + 51 | 0.50553 + 52 | 0.500825 + 18 | 0.496591 + 53 | 0.496208 + 54 | 0.491676 + 55 | 0.487225 + 56 | 0.482854 + 19 | 0.482412 + 57 | 0.47856 + 58 | 0.47434 + 59 | 0.470191 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306662 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +---------- + 0.999075 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 50 | 0.419425 + 49 | 0.417358 + 51 | 0.41482 + 52 | 0.410324 + 53 | 0.405932 + 54 | 0.40164 + 55 | 0.397443 + 56 | 0.393339 + 57 | 0.389323 + 58 | 0.385391 + 59 | 0.381541 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 50 | 0.481241 + 51 | 0.475957 + 52 | 0.470799 + 20 | 0.469285 + 53 | 0.465759 + 54 | 0.460835 + 21 | 0.457083 + 55 | 0.45602 + 4 | 0.45545 + 56 | 0.45131 + 57 | 0.446702 + 22 | 0.4457 + 58 | 0.442191 + 59 | 0.437774 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306545 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.413885 + 51 | 0.408856 + 52 | 0.403962 + 53 | 0.399199 + 54 | 0.394558 + 55 | 0.390035 + 56 | 0.385625 + 57 | 0.381322 + 58 | 0.377121 + 59 | 0.373019 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 50 | 0.425793 + 24 | 0.425047 + 51 | 0.420619 + 33 | 0.419273 + 25 | 0.415635 + 52 | 0.415585 + 34 | 0.41221 + 53 | 0.410684 + 26 | 0.406753 + 54 | 0.40591 + 35 | 0.405439 + 55 | 0.401257 + 36 | 0.398938 + 27 | 0.398352 + 56 | 0.39672 + 37 | 0.392689 + 57 | 0.392293 + 28 | 0.390389 + 49 | 0.390116 + 58 | 0.387972 + 38 | 0.386675 + 59 | 0.383752 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +CREATE INDEX idx_test_text ON test_text USING gist (v _text_sml_ops); +SET enable_seqscan = off; +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306753 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +---------- + 0.997241 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.398379 + 51 | 0.394635 + 52 | 0.390962 + 53 | 0.387358 + 54 | 0.38382 + 55 | 0.380346 + 56 | 0.376933 + 57 | 0.373581 + 58 | 0.370286 + 60 | 0.369012 + 59 | 0.367048 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 50 | 0.510326 + 51 | 0.50553 + 52 | 0.500825 + 18 | 0.496591 + 53 | 0.496208 + 54 | 0.491676 + 55 | 0.487225 + 56 | 0.482854 + 19 | 0.482412 + 57 | 0.47856 + 58 | 0.47434 + 59 | 0.470191 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306662 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +---------- + 0.999075 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 50 | 0.419425 + 49 | 0.417358 + 51 | 0.41482 + 52 | 0.410324 + 53 | 0.405932 + 54 | 0.40164 + 55 | 0.397443 + 56 | 0.393339 + 57 | 0.389323 + 58 | 0.385391 + 59 | 0.381541 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 50 | 0.481241 + 51 | 0.475957 + 52 | 0.470799 + 20 | 0.469285 + 53 | 0.465759 + 54 | 0.460835 + 21 | 0.457083 + 55 | 0.45602 + 4 | 0.45545 + 56 | 0.45131 + 57 | 0.446702 + 22 | 0.4457 + 58 | 0.442191 + 59 | 0.437774 + 23 | 0.435047 + 2 | 0.428355 + 24 | 0.425047 + 33 | 0.419273 + 25 | 0.415635 + 34 | 0.41221 + 26 | 0.406753 + 35 | 0.405439 + 36 | 0.398938 + 27 | 0.398352 + 37 | 0.392689 + 28 | 0.390389 + 49 | 0.390116 + 38 | 0.386675 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306545 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.413885 + 51 | 0.408856 + 52 | 0.403962 + 53 | 0.399199 + 54 | 0.394558 + 55 | 0.390035 + 56 | 0.385625 + 57 | 0.381322 + 58 | 0.377121 + 59 | 0.373019 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 50 | 0.425793 + 24 | 0.425047 + 51 | 0.420619 + 33 | 0.419273 + 25 | 0.415635 + 52 | 0.415585 + 34 | 0.41221 + 53 | 0.410684 + 26 | 0.406753 + 54 | 0.40591 + 35 | 0.405439 + 55 | 0.401257 + 36 | 0.398938 + 27 | 0.398352 + 56 | 0.39672 + 37 | 0.392689 + 57 | 0.392293 + 28 | 0.390389 + 49 | 0.390116 + 58 | 0.387972 + 38 | 0.386675 + 59 | 0.383752 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +DROP INDEX idx_test_text; +CREATE INDEX idx_test_text ON test_text USING gin (v _text_sml_ops); +SET enable_seqscan = off; +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +---------- + 0.997241 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +---------- + 0.999075 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; +ERROR: GIN supports only smlar.tf_method = "const" +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.870583 + 9 | 0.858716 + 12 | 0.766621 + 8 | 0.725108 + 13 | 0.681623 + 14 | 0.611081 + 7 | 0.599535 + 15 | 0.551777 + 16 | 0.501358 + 6 | 0.482432 + 17 | 0.458069 + 18 | 0.420577 + 19 | 0.387851 + 5 | 0.374337 + 20 | 0.359086 + 21 | 0.333644 + 22 | 0.311011 +(18 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 50 | 0.306545 + 49 | 0.306482 +(2 rows) + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 1 +(1 row) + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.951679 + 9 | 0.946916 + 12 | 0.909452 + 8 | 0.891104 + 13 | 0.872126 + 14 | 0.83881 + 7 | 0.83201 + 15 | 0.808827 + 16 | 0.781649 + 6 | 0.768873 + 17 | 0.756858 + 18 | 0.734118 + 19 | 0.713158 + 5 | 0.700594 + 20 | 0.693752 + 21 | 0.675714 + 22 | 0.658886 + 23 | 0.643138 + 24 | 0.628355 + 4 | 0.625485 + 25 | 0.61444 + 26 | 0.60131 + 27 | 0.58889 + 28 | 0.577119 + 29 | 0.565939 + 30 | 0.5553 + 31 | 0.54516 + 3 | 0.5407 + 32 | 0.535479 + 33 | 0.526223 + 34 | 0.517359 + 35 | 0.50886 + 36 | 0.5007 + 37 | 0.492857 + 38 | 0.48531 + 39 | 0.478039 + 40 | 0.471027 + 41 | 0.464259 + 42 | 0.45772 + 43 | 0.451396 + 44 | 0.445276 + 2 | 0.440679 + 45 | 0.439348 + 46 | 0.433601 + 47 | 0.428026 + 48 | 0.422615 + 49 | 0.417358 + 50 | 0.413885 + 51 | 0.408856 + 52 | 0.403962 + 53 | 0.399199 + 54 | 0.394558 + 55 | 0.390035 + 56 | 0.385625 + 57 | 0.381322 + 58 | 0.377121 + 59 | 0.373019 + 60 | 0.369012 + 61 | 0.365094 + 62 | 0.361263 + 63 | 0.357515 + 64 | 0.353847 + 65 | 0.350256 + 66 | 0.346738 + 67 | 0.343291 + 68 | 0.339913 + 69 | 0.336601 + 70 | 0.333351 + 71 | 0.330163 + 72 | 0.327034 + 73 | 0.323962 + 74 | 0.320945 + 75 | 0.317981 + 76 | 0.315068 + 77 | 0.312205 + 1 | 0.311044 + 78 | 0.30939 + 79 | 0.306621 + 80 | 0.303897 + 81 | 0.301217 +(81 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +-----+---------- + 8 | 0.759109 + 9 | 0.714367 + 7 | 0.694063 + 10 | 0.676445 + 11 | 0.643758 + 6 | 0.623275 + 12 | 0.615194 + 13 | 0.589945 + 14 | 0.567409 + 15 | 0.547127 + 5 | 0.544812 + 16 | 0.528742 + 17 | 0.511973 + 18 | 0.496591 + 19 | 0.482412 + 20 | 0.469285 + 21 | 0.457083 + 4 | 0.45545 + 22 | 0.4457 + 23 | 0.435047 + 2 | 0.428355 + 50 | 0.425793 + 24 | 0.425047 + 51 | 0.420619 + 33 | 0.419273 + 25 | 0.415635 + 52 | 0.415585 + 34 | 0.41221 + 53 | 0.410684 + 26 | 0.406753 + 54 | 0.40591 + 35 | 0.405439 + 55 | 0.401257 + 36 | 0.398938 + 27 | 0.398352 + 56 | 0.39672 + 37 | 0.392689 + 57 | 0.392293 + 28 | 0.390389 + 49 | 0.390116 + 58 | 0.387972 + 38 | 0.386675 + 59 | 0.383752 + 29 | 0.382826 + 39 | 0.380882 + 60 | 0.379628 + 30 | 0.37563 + 61 | 0.375598 + 40 | 0.375295 + 62 | 0.371657 + 41 | 0.369903 + 31 | 0.368771 + 63 | 0.367801 + 42 | 0.364692 + 64 | 0.364028 + 32 | 0.362222 + 65 | 0.360333 + 43 | 0.359654 + 66 | 0.356714 + 44 | 0.354778 + 67 | 0.353168 + 45 | 0.350054 + 68 | 0.349693 + 3 | 0.349116 + 69 | 0.346285 + 46 | 0.345476 + 70 | 0.342942 + 47 | 0.341034 + 71 | 0.339663 + 48 | 0.336722 + 72 | 0.336444 + 73 | 0.333283 + 74 | 0.330179 + 75 | 0.32713 + 76 | 0.324133 + 77 | 0.321188 + 78 | 0.318291 + 79 | 0.315443 + 80 | 0.312641 + 81 | 0.309883 + 82 | 0.307169 + 83 | 0.304497 + 1 | 0.302346 + 201 | 0.302346 + 84 | 0.301866 +(85 rows) + +RESET enable_seqscan; diff --git a/external/smlar/expected/texti.out b/external/smlar/expected/texti.out new file mode 100644 index 00000000000..b52ff2302f6 --- /dev/null +++ b/external/smlar/expected/texti.out @@ -0,0 +1,42 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(5.0); + set_smlar_limit +----------------- + 5 +(1 row) + +SET smlar.type='overlap'; +DROP INDEX idx_test_text; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); + smlar +------- + 2 +(1 row) + +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_text WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + t | s +-----+---- + 199 | 10 + 200 | 10 + 198 | 9 + 197 | 8 + 196 | 7 + 195 | 6 + 194 | 5 +(7 rows) + +CREATE INDEX idx_test_text ON test_text USING gin (v _text_sml_ops); +SET enable_seqscan = off; +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_text WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + t | s +-----+---- + 199 | 10 + 200 | 10 + 198 | 9 + 197 | 8 + 196 | 7 + 195 | 6 + 194 | 5 +(7 rows) + +RESET enable_seqscan; diff --git a/external/smlar/expected/time.out b/external/smlar/expected/time.out new file mode 100644 index 00000000000..e7247021abd --- /dev/null +++ b/external/smlar/expected/time.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + epoch2timestamp(v)::time + FROM + generate_series(1, t) as v + ) AS v + INTO test_time +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::time[]) AS s FROM test_time WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::time[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::time[]) AS s FROM test_time WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::time[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_time ON test_time USING gist (v _time_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::time[]) AS s FROM test_time WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::time[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::time[]) AS s FROM test_time WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::time[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_time; +CREATE INDEX idx_test_time ON test_time USING gin (v _time_sml_ops); +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::time[]) AS s FROM test_time WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::time[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::time[]) AS s FROM test_time WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::time[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/timestamp.out b/external/smlar/expected/timestamp.out new file mode 100644 index 00000000000..3ef57c25325 --- /dev/null +++ b/external/smlar/expected/timestamp.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + epoch2timestamp(v)::timestamp + FROM + generate_series(1, t) as v + ) AS v + INTO test_timestamp +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')) AS s FROM test_timestamp WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}') ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')) AS s FROM test_timestamp WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}') ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_timestamp ON test_timestamp USING gist (v _timestamp_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')) AS s FROM test_timestamp WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}') ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')) AS s FROM test_timestamp WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}') ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_timestamp; +CREATE INDEX idx_test_timestamp ON test_timestamp USING gin (v _timestamp_sml_ops); +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')) AS s FROM test_timestamp WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}') ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')) AS s FROM test_timestamp WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}') ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/timestamptz.out b/external/smlar/expected/timestamptz.out new file mode 100644 index 00000000000..a5dc8906abb --- /dev/null +++ b/external/smlar/expected/timestamptz.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + epoch2timestamp(v)::timestamptz + FROM + generate_series(1, t) as v + ) AS v + INTO test_timestamptz +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]) AS s FROM test_timestamptz WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]) AS s FROM test_timestamptz WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_timestamptz ON test_timestamptz USING gist (v _timestamptz_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]) AS s FROM test_timestamptz WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]) AS s FROM test_timestamptz WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_timestamptz; +CREATE INDEX idx_test_timestamptz ON test_timestamptz USING gin (v _timestamptz_sml_ops); +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]) AS s FROM test_timestamptz WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]) AS s FROM test_timestamptz WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/timetz.out b/external/smlar/expected/timetz.out new file mode 100644 index 00000000000..d0fa0e944a3 --- /dev/null +++ b/external/smlar/expected/timetz.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + epoch2timestamp(v)::timestamptz::timetz + FROM + generate_series(1, t) as v + ) AS v + INTO test_timetz +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]::timetz[]) AS s FROM test_timetz WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]::timetz[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]::timetz[]) AS s FROM test_timetz WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]::timetz[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_timetz ON test_timetz USING gist (v _timetz_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]::timetz[]) AS s FROM test_timetz WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]::timetz[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]::timetz[]) AS s FROM test_timetz WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]::timetz[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_timetz; +CREATE INDEX idx_test_timetz ON test_timetz USING gin (v _timetz_sml_ops); +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]::timetz[]) AS s FROM test_timetz WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]::timetz[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]::timetz[]) AS s FROM test_timetz WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]::timetz[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/varbit.out b/external/smlar/expected/varbit.out new file mode 100644 index 00000000000..6f614d824e1 --- /dev/null +++ b/external/smlar/expected/varbit.out @@ -0,0 +1,105 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::bit(10)::varbit + FROM + generate_series(1, t) as v + ) AS v + INTO test_varbit +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[]::varbit[]) AS s FROM test_varbit WHERE v % '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[]::varbit[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[]::varbit[]) AS s FROM test_varbit WHERE v % '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[]::varbit[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_varbit ON test_varbit USING gin (v _varbit_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[]::varbit[]) AS s FROM test_varbit WHERE v % '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[]::varbit[] ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[]::varbit[]) AS s FROM test_varbit WHERE v % '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[]::varbit[] ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/expected/varchar.out b/external/smlar/expected/varchar.out new file mode 100644 index 00000000000..f74b110d7ad --- /dev/null +++ b/external/smlar/expected/varchar.out @@ -0,0 +1,149 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + set_smlar_limit +----------------- + 0.6 +(1 row) + +SELECT + t, + ARRAY( + SELECT + v::varchar + FROM + generate_series(1, t) as v + ) AS v + INTO test_varchar +FROM + generate_series(1, 200) as t; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_varchar WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_varchar WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +CREATE INDEX idx_test_varchar ON test_varchar USING gist (v _varchar_sml_ops); +SET enable_seqscan=off; +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_varchar WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_varchar WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +DROP INDEX idx_test_varchar; +CREATE INDEX idx_test_varchar ON test_varchar USING gin (v _varchar_sml_ops); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_varchar WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 10 | 1 + 11 | 0.953463 + 9 | 0.948683 + 12 | 0.912871 + 8 | 0.894427 + 13 | 0.877058 + 14 | 0.845154 + 7 | 0.83666 + 15 | 0.816497 + 16 | 0.790569 + 6 | 0.774597 + 17 | 0.766965 + 18 | 0.745356 + 19 | 0.725476 + 5 | 0.707107 + 20 | 0.707107 + 21 | 0.690066 + 22 | 0.6742 + 23 | 0.65938 + 24 | 0.645497 + 4 | 0.632456 + 25 | 0.632456 + 26 | 0.620174 + 27 | 0.608581 +(24 rows) + +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_varchar WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + t | s +----+---------- + 8 | 0.782624 + 9 | 0.737865 + 7 | 0.717137 + 10 | 0.7 + 11 | 0.667424 + 6 | 0.645497 + 12 | 0.63901 + 13 | 0.613941 +(8 rows) + +SET enable_seqscan=on; diff --git a/external/smlar/smlar--1.0.sql b/external/smlar/smlar--1.0.sql new file mode 100644 index 00000000000..fcce8146082 --- /dev/null +++ b/external/smlar/smlar--1.0.sql @@ -0,0 +1,714 @@ +-- complain if script is sourced in psql, rather than via CREATE EXTENSION +\echo Use "CREATE EXTENSION smlar" to load this file. \quit + +CREATE OR REPLACE FUNCTION smlar(anyarray, anyarray) + RETURNS float4 + AS 'MODULE_PATHNAME', 'arraysml' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION smlar(anyarray, anyarray, bool) + RETURNS float4 + AS 'MODULE_PATHNAME', 'arraysmlw' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION smlar(anyarray, anyarray, text) + RETURNS float4 + AS 'MODULE_PATHNAME', 'arraysml_func' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION set_smlar_limit(float4) + RETURNS float4 + AS 'MODULE_PATHNAME' + LANGUAGE C STRICT VOLATILE; + +CREATE OR REPLACE FUNCTION show_smlar_limit() + RETURNS float4 + AS 'MODULE_PATHNAME' + LANGUAGE C STRICT STABLE; + +CREATE OR REPLACE FUNCTION smlar_op(anyarray,anyarray) + RETURNS bool + AS 'MODULE_PATHNAME', 'arraysml_op' + LANGUAGE C STRICT STABLE; + +CREATE OPERATOR % ( + LEFTARG = anyarray, + RIGHTARG = anyarray, + PROCEDURE = smlar_op, + COMMUTATOR = '%', + RESTRICT = contsel, + JOIN = contjoinsel +); + +-- + +CREATE OR REPLACE FUNCTION tsvector2textarray(tsvector) + RETURNS _text + AS 'MODULE_PATHNAME', 'tsvector2textarray' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION array_unique(anyarray) + RETURNS anyarray + AS 'MODULE_PATHNAME', 'array_unique' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION inarray(anyarray, anyelement) + RETURNS float4 + AS 'MODULE_PATHNAME', 'inarray' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION inarray(anyarray, anyelement, float4, float4) + RETURNS float4 + AS 'MODULE_PATHNAME', 'inarray' + LANGUAGE C STRICT IMMUTABLE; + +--gist key + +CREATE OR REPLACE FUNCTION gsmlsign_in(cstring) + RETURNS gsmlsign + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_out(gsmlsign) + RETURNS cstring + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE TYPE gsmlsign ( + INTERNALLENGTH = -1, + INPUT = gsmlsign_in, + OUTPUT = gsmlsign_out +); + +-- support functions for gist +CREATE OR REPLACE FUNCTION gsmlsign_consistent(gsmlsign,internal,int4) + RETURNS bool + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_compress(internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_decompress(internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_penalty(internal,internal,internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_picksplit(internal, internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_union(bytea, internal) + RETURNS _int4 + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_same(gsmlsign, gsmlsign, internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +--create the operator classes for gist + +CREATE OPERATOR CLASS _int2_sml_ops +FOR TYPE _int2 USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _int4_sml_ops +FOR TYPE _int4 USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _int8_sml_ops +FOR TYPE _int8 USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _float4_sml_ops +FOR TYPE _float4 USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _float8_sml_ops +FOR TYPE _float8 USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +-- money type hasn't hash support +--CREATE OPERATOR CLASS _money_sml_ops +--FOR TYPE _money USING gist +--AS +-- OPERATOR 1 && (anyarray, anyarray), +-- OPERATOR 2 % (anyarray, anyarray), +-- FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), +-- FUNCTION 2 gsmlsign_union (bytea, internal), +-- FUNCTION 3 gsmlsign_compress (internal), +-- FUNCTION 4 gsmlsign_decompress (internal), +-- FUNCTION 5 gsmlsign_penalty (internal, internal, internal), +-- FUNCTION 6 gsmlsign_picksplit (internal, internal), +-- FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +--STORAGE gsmlsign; + +CREATE OPERATOR CLASS _oid_sml_ops +FOR TYPE _oid USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _timestamp_sml_ops +FOR TYPE _timestamp USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _timestamptz_sml_ops +FOR TYPE _timestamptz USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _time_sml_ops +FOR TYPE _time USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _timetz_sml_ops +FOR TYPE _timetz USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _date_sml_ops +FOR TYPE _date USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _interval_sml_ops +FOR TYPE _interval USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _macaddr_sml_ops +FOR TYPE _macaddr USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _inet_sml_ops +FOR TYPE _inet USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _cidr_sml_ops +FOR TYPE _cidr USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _text_sml_ops +FOR TYPE _text USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _varchar_sml_ops +FOR TYPE _varchar USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _char_sml_ops +FOR TYPE _char USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _bytea_sml_ops +FOR TYPE _bytea USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +--no hash +--CREATE OPERATOR CLASS _bit_sml_ops +--FOR TYPE _bit USING gist +--AS +-- OPERATOR 1 && (anyarray, anyarray), +-- OPERATOR 2 % (anyarray, anyarray), +-- FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), +-- FUNCTION 2 gsmlsign_union (bytea, internal), +-- FUNCTION 3 gsmlsign_compress (internal), +-- FUNCTION 4 gsmlsign_decompress (internal), +-- FUNCTION 5 gsmlsign_penalty (internal, internal, internal), +-- FUNCTION 6 gsmlsign_picksplit (internal, internal), +-- FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +--STORAGE gsmlsign; + +--CREATE OPERATOR CLASS _varbit_sml_ops +--FOR TYPE _varbit USING gist +--AS +-- OPERATOR 1 && (anyarray, anyarray), +-- OPERATOR 2 % (anyarray, anyarray), +-- FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), +-- FUNCTION 2 gsmlsign_union (bytea, internal), +-- FUNCTION 3 gsmlsign_compress (internal), +-- FUNCTION 4 gsmlsign_decompress (internal), +-- FUNCTION 5 gsmlsign_penalty (internal, internal, internal), +-- FUNCTION 6 gsmlsign_picksplit (internal, internal), +-- FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +--STORAGE gsmlsign; + +CREATE OPERATOR CLASS _numeric_sml_ops +FOR TYPE _numeric USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +--gin support functions +CREATE OR REPLACE FUNCTION smlararrayextract(anyarray, internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION smlarqueryarrayextract(anyarray, internal, internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION smlararrayconsistent(internal, internal, anyarray) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +--gin opclasses + +CREATE OPERATOR CLASS _int2_sml_ops +FOR TYPE _int2 USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btint2cmp(int2,int2), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE int2; +CREATE OPERATOR CLASS _int4_sml_ops +FOR TYPE _int4 USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btint4cmp(int4,int4), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE int4; + +CREATE OPERATOR CLASS _int8_sml_ops +FOR TYPE _int8 USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btint8cmp(int8,int8), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE int8; + +CREATE OPERATOR CLASS _float4_sml_ops +FOR TYPE _float4 USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btfloat4cmp(float4,float4), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE float4; + +CREATE OPERATOR CLASS _float8_sml_ops +FOR TYPE _float8 USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btfloat8cmp(float8,float8), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE float8; + +CREATE OPERATOR CLASS _money_sml_ops +FOR TYPE _money USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 cash_cmp(money,money), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE money; + +CREATE OPERATOR CLASS _oid_sml_ops +FOR TYPE _oid USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btoidcmp(oid,oid), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE oid; + +CREATE OPERATOR CLASS _timestamp_sml_ops +FOR TYPE _timestamp USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 timestamp_cmp(timestamp,timestamp), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE timestamp; + +CREATE OPERATOR CLASS _timestamptz_sml_ops +FOR TYPE _timestamptz USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 timestamptz_cmp(timestamptz,timestamptz), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE timestamptz; + +CREATE OPERATOR CLASS _time_sml_ops +FOR TYPE _time USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 time_cmp(time,time), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE time; + +CREATE OPERATOR CLASS _timetz_sml_ops +FOR TYPE _timetz USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 timetz_cmp(timetz,timetz), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE timetz; + +CREATE OPERATOR CLASS _date_sml_ops +FOR TYPE _date USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 date_cmp(date,date), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE date; + +CREATE OPERATOR CLASS _interval_sml_ops +FOR TYPE _interval USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 interval_cmp(interval,interval), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE interval; + +CREATE OPERATOR CLASS _macaddr_sml_ops +FOR TYPE _macaddr USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 macaddr_cmp(macaddr,macaddr), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE macaddr; + +CREATE OPERATOR CLASS _inet_sml_ops +FOR TYPE _inet USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 network_cmp(inet,inet), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE inet; + +CREATE OPERATOR CLASS _cidr_sml_ops +FOR TYPE _cidr USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 network_cmp(inet,inet), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE cidr; + +CREATE OPERATOR CLASS _text_sml_ops +FOR TYPE _text USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 bttextcmp(text,text), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE text; + +CREATE OPERATOR CLASS _varchar_sml_ops +FOR TYPE _varchar USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 bttextcmp(text,text), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE varchar; + +CREATE OPERATOR CLASS _char_sml_ops +FOR TYPE "_char" USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btcharcmp("char","char"), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE "char"; + +CREATE OPERATOR CLASS _bytea_sml_ops +FOR TYPE _bytea USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 byteacmp(bytea,bytea), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE bytea; + +CREATE OPERATOR CLASS _bit_sml_ops +FOR TYPE _bit USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 bitcmp(bit,bit), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE bit; + +CREATE OPERATOR CLASS _varbit_sml_ops +FOR TYPE _varbit USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 varbitcmp(varbit,varbit), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE varbit; + +CREATE OPERATOR CLASS _numeric_sml_ops +FOR TYPE _numeric USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 numeric_cmp(numeric,numeric), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE numeric; + diff --git a/external/smlar/smlar--unpackaged--1.0.sql b/external/smlar/smlar--unpackaged--1.0.sql new file mode 100644 index 00000000000..36f8720b3eb --- /dev/null +++ b/external/smlar/smlar--unpackaged--1.0.sql @@ -0,0 +1,125 @@ +-- complain if script is sourced in psql, rather than via CREATE EXTENSION +\echo Use "CREATE EXTENSION extension_name" to load this file. \quit + +ALTER EXTENSION smlar ADD FUNCTION tsvector2textarray(tsvector); +ALTER EXTENSION smlar ADD FUNCTION array_unique(anyarray); +ALTER EXTENSION smlar ADD FUNCTION inarray(anyarray, anyelement); +ALTER EXTENSION smlar ADD FUNCTION inarray(anyarray, anyelement, float4, float4); + +ALTER EXTENSION smlar ADD FUNCTION smlar(anyarray, anyarray); +ALTER EXTENSION smlar ADD FUNCTION smlar(anyarray, anyarray, bool); +ALTER EXTENSION smlar ADD FUNCTION smlar(anyarray, anyarray, text); +ALTER EXTENSION smlar ADD FUNCTION set_smlar_limit(float4); +ALTER EXTENSION smlar ADD FUNCTION show_smlar_limit(); +ALTER EXTENSION smlar ADD FUNCTION smlar_op(anyarray,anyarray); +ALTER EXTENSION smlar ADD OPERATOR % (anyarray,anyarray); + +ALTER EXTENSION smlar ADD FUNCTION gsmlsign_out(gsmlsign); +ALTER EXTENSION smlar ADD FUNCTION gsmlsign_in(cstring); +ALTER EXTENSION smlar ADD TYPE gsmlsign; + +ALTER EXTENSION smlar ADD FUNCTION gsmlsign_consistent(gsmlsign,internal,int4); +ALTER EXTENSION smlar ADD FUNCTION gsmlsign_compress(internal); +ALTER EXTENSION smlar ADD FUNCTION gsmlsign_decompress(internal); +ALTER EXTENSION smlar ADD FUNCTION gsmlsign_penalty(internal,internal,internal); +ALTER EXTENSION smlar ADD FUNCTION gsmlsign_picksplit(internal, internal); +ALTER EXTENSION smlar ADD FUNCTION gsmlsign_union(bytea, internal); +ALTER EXTENSION smlar ADD FUNCTION gsmlsign_same(gsmlsign, gsmlsign, internal); + +ALTER EXTENSION smlar ADD FUNCTION smlararrayextract(anyarray, internal); +ALTER EXTENSION smlar ADD FUNCTION smlarqueryarrayextract(anyarray, internal, internal); +ALTER EXTENSION smlar ADD FUNCTION smlararrayconsistent(internal, internal, anyarray); + +ALTER EXTENSION smlar ADD OPERATOR CLASS _int2_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _int4_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _int8_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _float4_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _float8_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _money_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _oid_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _timestamp_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _timestamptz_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _time_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _timetz_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _date_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _interval_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _macaddr_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _inet_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _cidr_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _text_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _varchar_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _char_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _bytea_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _bit_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _varbit_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR CLASS _numeric_sml_ops USING gin; + +ALTER EXTENSION smlar ADD OPERATOR FAMILY _int2_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _int4_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _int8_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _float4_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _float8_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _money_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _oid_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _timestamp_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _timestamptz_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _time_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _timetz_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _date_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _interval_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _macaddr_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _inet_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _cidr_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _text_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _varchar_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _char_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _bytea_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _bit_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _varbit_sml_ops USING gin; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _numeric_sml_ops USING gin; + +ALTER EXTENSION smlar ADD OPERATOR CLASS _int2_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _int4_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _int8_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _float4_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _float8_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _oid_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _timestamp_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _timestamptz_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _time_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _timetz_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _date_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _interval_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _macaddr_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _inet_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _cidr_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _text_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _varchar_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _char_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _bytea_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR CLASS _numeric_sml_ops USING gist; + +ALTER EXTENSION smlar ADD OPERATOR FAMILY _int2_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _int4_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _int8_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _float4_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _float8_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _oid_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _timestamp_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _timestamptz_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _time_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _timetz_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _date_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _interval_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _macaddr_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _inet_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _cidr_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _text_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _varchar_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _char_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _bytea_sml_ops USING gist; +ALTER EXTENSION smlar ADD OPERATOR FAMILY _numeric_sml_ops USING gist; + + + + diff --git a/external/smlar/smlar.c b/external/smlar/smlar.c new file mode 100644 index 00000000000..ca0651f5e01 --- /dev/null +++ b/external/smlar/smlar.c @@ -0,0 +1,1021 @@ +#include "smlar.h" + +#include "fmgr.h" +#include "access/genam.h" +#include "access/heapam.h" +#include "access/htup_details.h" +#include "access/nbtree.h" +#include "catalog/indexing.h" +#include "catalog/pg_am.h" +#include "catalog/pg_amproc.h" +#include "catalog/pg_cast.h" +#include "catalog/pg_opclass.h" +#include "catalog/pg_type.h" +#include "executor/spi.h" +#include "utils/catcache.h" +#include "utils/fmgroids.h" +#include "utils/lsyscache.h" +#include "utils/memutils.h" +#if (PG_VERSION_NUM < 120000) +#include "utils/tqual.h" +#endif +#include "utils/syscache.h" +#include "utils/typcache.h" + +PG_MODULE_MAGIC; + +#if (PG_VERSION_NUM >= 90400) +#define SNAPSHOT NULL +#else +#define SNAPSHOT SnapshotNow +#endif + +#if PG_VERSION_NUM >= 130000 +#define heap_open(r, l) table_open((r), (l)) +#define heap_close(r, l) table_close((r), (l)) +#endif + +static Oid +getDefaultOpclass(Oid amoid, Oid typid) +{ + ScanKeyData skey; + SysScanDesc scan; + HeapTuple tuple; + Relation heapRel; + Oid opclassOid = InvalidOid; + + heapRel = heap_open(OperatorClassRelationId, AccessShareLock); + + ScanKeyInit(&skey, + Anum_pg_opclass_opcmethod, + BTEqualStrategyNumber, F_OIDEQ, + ObjectIdGetDatum(amoid)); + + scan = systable_beginscan(heapRel, + OpclassAmNameNspIndexId, true, + SNAPSHOT, 1, &skey); + + while (HeapTupleIsValid((tuple = systable_getnext(scan)))) + { + Form_pg_opclass opclass = (Form_pg_opclass)GETSTRUCT(tuple); + + if ( opclass->opcintype == typid && opclass->opcdefault ) + { + if ( OidIsValid(opclassOid) ) + elog(ERROR, "Ambiguous opclass for type %u (access method %u)", typid, amoid); +#if (PG_VERSION_NUM >= 120000) + opclassOid = opclass->oid; +#else + opclassOid = HeapTupleGetOid(tuple); +#endif + } + } + + systable_endscan(scan); + heap_close(heapRel, AccessShareLock); + + return opclassOid; +} + +static Oid +getAMProc(Oid amoid, Oid typid) +{ + Oid opclassOid = getDefaultOpclass(amoid, typid); + Oid procOid = InvalidOid; + Oid opfamilyOid; + ScanKeyData skey[4]; + SysScanDesc scan; + HeapTuple tuple; + Relation heapRel; + + if ( !OidIsValid(opclassOid) ) + { + typid = getBaseType(typid); + opclassOid = getDefaultOpclass(amoid, typid); + } + + if ( !OidIsValid(opclassOid) ) + { + CatCList *catlist; + int i; + + /* + * Search binary-coercible type + */ +#ifdef SearchSysCacheList1 + catlist = SearchSysCacheList1(CASTSOURCETARGET, + ObjectIdGetDatum(typid)); +#else + catlist = SearchSysCacheList(CASTSOURCETARGET, 1, + ObjectIdGetDatum(typid), + 0, 0, 0); +#endif + + for (i = 0; i < catlist->n_members; i++) + { + HeapTuple tuple = &catlist->members[i]->tuple; + Form_pg_cast castForm = (Form_pg_cast)GETSTRUCT(tuple); + + if ( castForm->castfunc == InvalidOid && castForm->castcontext == COERCION_CODE_IMPLICIT ) + { + typid = castForm->casttarget; + opclassOid = getDefaultOpclass(amoid, typid); + if( OidIsValid(opclassOid) ) + break; + } + } + + ReleaseSysCacheList(catlist); + } + + if ( !OidIsValid(opclassOid) ) + return InvalidOid; + + opfamilyOid = get_opclass_family(opclassOid); + + heapRel = heap_open(AccessMethodProcedureRelationId, AccessShareLock); + ScanKeyInit(&skey[0], + Anum_pg_amproc_amprocfamily, + BTEqualStrategyNumber, F_OIDEQ, + ObjectIdGetDatum(opfamilyOid)); + ScanKeyInit(&skey[1], + Anum_pg_amproc_amproclefttype, + BTEqualStrategyNumber, F_OIDEQ, + ObjectIdGetDatum(typid)); + ScanKeyInit(&skey[2], + Anum_pg_amproc_amprocrighttype, + BTEqualStrategyNumber, F_OIDEQ, + ObjectIdGetDatum(typid)); +#if PG_VERSION_NUM >= 90200 + ScanKeyInit(&skey[3], + Anum_pg_amproc_amprocnum, + BTEqualStrategyNumber, F_OIDEQ, + Int32GetDatum(BTORDER_PROC)); +#endif + + scan = systable_beginscan(heapRel, AccessMethodProcedureIndexId, true, + SNAPSHOT, +#if PG_VERSION_NUM >= 90200 + 4, +#else + 3, +#endif + skey); + while (HeapTupleIsValid(tuple = systable_getnext(scan))) + { + Form_pg_amproc amprocform = (Form_pg_amproc) GETSTRUCT(tuple); + + switch(amoid) + { + case BTREE_AM_OID: + case HASH_AM_OID: + if ( OidIsValid(procOid) ) + elog(ERROR,"Ambiguous support function for type %u (opclass %u)", typid, opfamilyOid); + procOid = amprocform->amproc; + break; + default: + elog(ERROR,"Unsupported access method"); + } + } + + systable_endscan(scan); + heap_close(heapRel, AccessShareLock); + + return procOid; +} + +static ProcTypeInfo *cacheProcs = NULL; +static int nCacheProcs = 0; + +#ifndef TupleDescAttr +#define TupleDescAttr(tupdesc, i) ((tupdesc)->attrs[(i)]) +#endif + +static ProcTypeInfo +fillProcs(Oid typid) +{ + ProcTypeInfo info = malloc(sizeof(ProcTypeInfoData)); + + if (!info) + elog(ERROR, "Can't allocate %u memory", (uint32)sizeof(ProcTypeInfoData)); + + info->typid = typid; + info->typtype = get_typtype(typid); + + if (info->typtype == 'c') + { + /* composite type */ + TupleDesc tupdesc; + MemoryContext oldcontext; + + tupdesc = lookup_rowtype_tupdesc(typid, -1); + + if (tupdesc->natts != 2) + elog(ERROR,"Composite type has wrong number of fields"); + if (TupleDescAttr(tupdesc, 1)->atttypid != FLOAT4OID) + elog(ERROR,"Second field of composite type is not float4"); + + oldcontext = MemoryContextSwitchTo(TopMemoryContext); + info->tupDesc = CreateTupleDescCopyConstr(tupdesc); + MemoryContextSwitchTo(oldcontext); + + ReleaseTupleDesc(tupdesc); + + info->cmpFuncOid = getAMProc(BTREE_AM_OID, + TupleDescAttr(info->tupDesc, 0)->atttypid); + info->hashFuncOid = getAMProc(HASH_AM_OID, + TupleDescAttr(info->tupDesc, 0)->atttypid); + } + else + { + info->tupDesc = NULL; + + /* plain type */ + info->cmpFuncOid = getAMProc(BTREE_AM_OID, typid); + info->hashFuncOid = getAMProc(HASH_AM_OID, typid); + } + + get_typlenbyvalalign(typid, &info->typlen, &info->typbyval, &info->typalign); + info->hashFuncInited = info->cmpFuncInited = false; + + + return info; +} + +void +getFmgrInfoCmp(ProcTypeInfo info) +{ + if ( info->cmpFuncInited == false ) + { + if ( !OidIsValid(info->cmpFuncOid) ) + elog(ERROR, "Could not find cmp function for type %u", info->typid); + + fmgr_info_cxt( info->cmpFuncOid, &info->cmpFunc, TopMemoryContext ); + info->cmpFuncInited = true; + } +} + +void +getFmgrInfoHash(ProcTypeInfo info) +{ + if ( info->hashFuncInited == false ) + { + if ( !OidIsValid(info->hashFuncOid) ) + elog(ERROR, "Could not find hash function for type %u", info->typid); + + fmgr_info_cxt( info->hashFuncOid, &info->hashFunc, TopMemoryContext ); + info->hashFuncInited = true; + } +} + +static int +cmpProcTypeInfo(const void *a, const void *b) +{ + ProcTypeInfo av = *(ProcTypeInfo*)a; + ProcTypeInfo bv = *(ProcTypeInfo*)b; + + Assert( av->typid != bv->typid ); + + return ( av->typid > bv->typid ) ? 1 : -1; +} + +ProcTypeInfo +findProcs(Oid typid) +{ + ProcTypeInfo info = NULL; + + if ( nCacheProcs == 1 ) + { + if ( cacheProcs[0]->typid == typid ) + { + /*cacheProcs[0]->hashFuncInited = cacheProcs[0]->cmpFuncInited = false;*/ + return cacheProcs[0]; + } + } + else if ( nCacheProcs > 1 ) + { + ProcTypeInfo *StopMiddle; + ProcTypeInfo *StopLow = cacheProcs, + *StopHigh = cacheProcs + nCacheProcs; + + while (StopLow < StopHigh) { + StopMiddle = StopLow + ((StopHigh - StopLow) >> 1); + info = *StopMiddle; + + if ( info->typid == typid ) + { + /* info->hashFuncInited = info->cmpFuncInited = false; */ + return info; + } + else if ( info->typid < typid ) + StopLow = StopMiddle + 1; + else + StopHigh = StopMiddle; + } + + /* not found */ + } + + info = fillProcs(typid); + if ( nCacheProcs == 0 ) + { + cacheProcs = malloc(sizeof(ProcTypeInfo)); + + if (!cacheProcs) + elog(ERROR, "Can't allocate %u memory", (uint32)sizeof(ProcTypeInfo)); + else + { + nCacheProcs = 1; + cacheProcs[0] = info; + } + } + else + { + ProcTypeInfo *cacheProcsTmp = realloc(cacheProcs, (nCacheProcs+1) * sizeof(ProcTypeInfo)); + + if (!cacheProcsTmp) + elog(ERROR, "Can't allocate %u memory", (uint32)sizeof(ProcTypeInfo) * (nCacheProcs+1)); + else + { + cacheProcs = cacheProcsTmp; + cacheProcs[ nCacheProcs ] = info; + nCacheProcs++; + qsort(cacheProcs, nCacheProcs, sizeof(ProcTypeInfo), cmpProcTypeInfo); + } + } + + /* info->hashFuncInited = info->cmpFuncInited = false; */ + + return info; +} + +/* + * WARNING. Array2SimpleArray* doesn't copy Datum! + */ +SimpleArray * +Array2SimpleArray(ProcTypeInfo info, ArrayType *a) +{ + SimpleArray *s = palloc(sizeof(SimpleArray)); + + CHECKARRVALID(a); + + if ( info == NULL ) + info = findProcs(ARR_ELEMTYPE(a)); + + s->info = info; + s->df = NULL; + s->hash = NULL; + + deconstruct_array(a, info->typid, + info->typlen, info->typbyval, info->typalign, + &s->elems, NULL, &s->nelems); + + return s; +} + +static Datum +deconstructCompositeType(ProcTypeInfo info, Datum in, double *weight) +{ + HeapTupleHeader rec = DatumGetHeapTupleHeader(in); + HeapTupleData tuple; + Datum values[2]; + bool nulls[2]; + + /* Build a temporary HeapTuple control structure */ + tuple.t_len = HeapTupleHeaderGetDatumLength(rec); + ItemPointerSetInvalid(&(tuple.t_self)); + tuple.t_tableOid = InvalidOid; + tuple.t_data = rec; + + heap_deform_tuple(&tuple, info->tupDesc, values, nulls); + if (nulls[0] || nulls[1]) + elog(ERROR, "Both fields in composite type could not be NULL"); + + if (weight) + *weight = DatumGetFloat4(values[1]); + return values[0]; +} + +static int +cmpArrayElem(const void *a, const void *b, void *arg) +{ + ProcTypeInfo info = (ProcTypeInfo)arg; + + if (info->tupDesc) + /* composite type */ + return DatumGetInt32( FCall2( &info->cmpFunc, + deconstructCompositeType(info, *(Datum*)a, NULL), + deconstructCompositeType(info, *(Datum*)b, NULL) ) ); + + return DatumGetInt32( FCall2( &info->cmpFunc, + *(Datum*)a, *(Datum*)b ) ); +} + +SimpleArray * +Array2SimpleArrayS(ProcTypeInfo info, ArrayType *a) +{ + SimpleArray *s = Array2SimpleArray(info, a); + + if ( s->nelems > 1 ) + { + getFmgrInfoCmp(s->info); + + qsort_arg(s->elems, s->nelems, sizeof(Datum), cmpArrayElem, s->info); + } + + return s; +} + +typedef struct cmpArrayElemData { + ProcTypeInfo info; + bool hasDuplicate; + +} cmpArrayElemData; + +static int +cmpArrayElemArg(const void *a, const void *b, void *arg) +{ + cmpArrayElemData *data = (cmpArrayElemData*)arg; + int res; + + if (data->info->tupDesc) + res = DatumGetInt32( FCall2( &data->info->cmpFunc, + deconstructCompositeType(data->info, *(Datum*)a, NULL), + deconstructCompositeType(data->info, *(Datum*)b, NULL) ) ); + else + res = DatumGetInt32( FCall2( &data->info->cmpFunc, + *(Datum*)a, *(Datum*)b ) ); + + if ( res == 0 ) + data->hasDuplicate = true; + + return res; +} + +/* + * Uniquefy array and calculate TF. Although + * result doesn't depend on normalization, we + * normalize TF by length array to have possiblity + * to limit estimation for index support. + * + * Cache signals of needing of TF caclulation + */ + +SimpleArray * +Array2SimpleArrayU(ProcTypeInfo info, ArrayType *a, void *cache) +{ + SimpleArray *s = Array2SimpleArray(info, a); + StatElem *stat = NULL; + + if ( s->nelems > 0 && cache ) + { + s->df = palloc(sizeof(double) * s->nelems); + s->df[0] = 1.0; /* init */ + } + + if ( s->nelems > 1 ) + { + cmpArrayElemData data; + int i; + + getFmgrInfoCmp(s->info); + data.info = s->info; + data.hasDuplicate = false; + + qsort_arg(s->elems, s->nelems, sizeof(Datum), cmpArrayElemArg, &data); + + if ( data.hasDuplicate ) + { + Datum *tmp, + *dr, + *data; + int num = s->nelems, + cmp; + + data = tmp = dr = s->elems; + + while (tmp - data < num) + { + cmp = (tmp == dr) ? 0 : cmpArrayElem(tmp, dr, s->info); + if ( cmp != 0 ) + { + *(++dr) = *tmp++; + if ( cache ) + s->df[ dr - data ] = 1.0; + } + else + { + if ( cache ) + s->df[ dr - data ] += 1.0; + tmp++; + } + } + + s->nelems = dr + 1 - s->elems; + + if ( cache ) + { + int tfm = getTFMethod(); + + for(i=0;inelems;i++) + { + stat = fingArrayStat(cache, s->info->typid, s->elems[i], stat); + if ( stat ) + { + switch(tfm) + { + case TF_LOG: + s->df[i] = (1.0 + log( s->df[i] )); + /* FALLTHROUGH */ + case TF_N: + s->df[i] *= stat->idf; + break; + case TF_CONST: + s->df[i] = stat->idf; + break; + default: + elog(ERROR,"Unknown TF method: %d", tfm); + } + } + else + { + s->df[i] = 0.0; /* unknown word */ + } + } + } + } + else if ( cache ) + { + for(i=0;inelems;i++) + { + stat = fingArrayStat(cache, s->info->typid, s->elems[i], stat); + if ( stat ) + s->df[i] = stat->idf; + else + s->df[i] = 0.0; + } + } + } + else if (s->nelems > 0 && cache) + { + stat = fingArrayStat(cache, s->info->typid, s->elems[0], stat); + if ( stat ) + s->df[0] = stat->idf; + else + s->df[0] = 0.0; + } + + return s; +} + +static int +numOfIntersect(SimpleArray *a, SimpleArray *b) +{ + int cnt = 0, + cmp; + Datum *aptr = a->elems, + *bptr = b->elems; + ProcTypeInfo info = a->info; + + Assert( a->info->typid == b->info->typid ); + + getFmgrInfoCmp(info); + + while( aptr - a->elems < a->nelems && bptr - b->elems < b->nelems ) + { + cmp = cmpArrayElem(aptr, bptr, info); + if ( cmp < 0 ) + aptr++; + else if ( cmp > 0 ) + bptr++; + else + { + cnt++; + aptr++; + bptr++; + } + } + + return cnt; +} + +static double +TFIDFSml(SimpleArray *a, SimpleArray *b) +{ + int cmp; + Datum *aptr = a->elems, + *bptr = b->elems; + ProcTypeInfo info = a->info; + double res = 0.0; + double suma = 0.0, sumb = 0.0; + + Assert( a->info->typid == b->info->typid ); + Assert( a->df ); + Assert( b->df ); + + getFmgrInfoCmp(info); + + while( aptr - a->elems < a->nelems && bptr - b->elems < b->nelems ) + { + cmp = cmpArrayElem(aptr, bptr, info); + if ( cmp < 0 ) + { + suma += a->df[ aptr - a->elems ] * a->df[ aptr - a->elems ]; + aptr++; + } + else if ( cmp > 0 ) + { + sumb += b->df[ bptr - b->elems ] * b->df[ bptr - b->elems ]; + bptr++; + } + else + { + res += a->df[ aptr - a->elems ] * b->df[ bptr - b->elems ]; + suma += a->df[ aptr - a->elems ] * a->df[ aptr - a->elems ]; + sumb += b->df[ bptr - b->elems ] * b->df[ bptr - b->elems ]; + aptr++; + bptr++; + } + } + + /* + * Compute last elements + */ + while( aptr - a->elems < a->nelems ) + { + suma += a->df[ aptr - a->elems ] * a->df[ aptr - a->elems ]; + aptr++; + } + + while( bptr - b->elems < b->nelems ) + { + sumb += b->df[ bptr - b->elems ] * b->df[ bptr - b->elems ]; + bptr++; + } + + if ( suma > 0.0 && sumb > 0.0 ) + res = res / sqrt( suma * sumb ); + else + res = 0.0; + + return res; +} + + +PG_FUNCTION_INFO_V1(arraysml); +Datum arraysml(PG_FUNCTION_ARGS); +Datum +arraysml(PG_FUNCTION_ARGS) +{ + ArrayType *a, *b; + SimpleArray *sa, *sb; + + fcinfo->flinfo->fn_extra = SearchArrayCache( + fcinfo->flinfo->fn_extra, + fcinfo->flinfo->fn_mcxt, + PG_GETARG_DATUM(0), &a, &sa, NULL); + fcinfo->flinfo->fn_extra = SearchArrayCache( + fcinfo->flinfo->fn_extra, + fcinfo->flinfo->fn_mcxt, + PG_GETARG_DATUM(1), &b, &sb, NULL); + + if ( ARR_ELEMTYPE(a) != ARR_ELEMTYPE(b) ) + elog(ERROR,"Arguments array are not the same type!"); + + if (ARRISVOID(a) || ARRISVOID(b)) + PG_RETURN_FLOAT4(0.0); + + switch(getSmlType()) + { + case ST_TFIDF: + PG_RETURN_FLOAT4( TFIDFSml(sa, sb) ); + break; + case ST_COSINE: + { + int cnt; + double power; + + power = ((double)(sa->nelems)) * ((double)(sb->nelems)); + cnt = numOfIntersect(sa, sb); + + PG_RETURN_FLOAT4( ((double)cnt) / sqrt( power ) ); + } + break; + case ST_OVERLAP: + { + float4 res = (float4)numOfIntersect(sa, sb); + + PG_RETURN_FLOAT4(res); + } + break; + default: + elog(ERROR,"Unsupported formula type of similarity"); + } + + PG_RETURN_FLOAT4(0.0); /* keep compiler quiet */ +} + +PG_FUNCTION_INFO_V1(arraysmlw); +Datum arraysmlw(PG_FUNCTION_ARGS); +Datum +arraysmlw(PG_FUNCTION_ARGS) +{ + ArrayType *a, *b; + SimpleArray *sa, *sb; + bool useIntersect = PG_GETARG_BOOL(2); + double numerator = 0.0; + double denominatorA = 0.0, + denominatorB = 0.0, + tmpA, tmpB; + int cmp; + ProcTypeInfo info; + int ai = 0, bi = 0; + + fcinfo->flinfo->fn_extra = SearchArrayCache( + fcinfo->flinfo->fn_extra, + fcinfo->flinfo->fn_mcxt, + PG_GETARG_DATUM(0), &a, &sa, NULL); + fcinfo->flinfo->fn_extra = SearchArrayCache( + fcinfo->flinfo->fn_extra, + fcinfo->flinfo->fn_mcxt, + PG_GETARG_DATUM(1), &b, &sb, NULL); + + if ( ARR_ELEMTYPE(a) != ARR_ELEMTYPE(b) ) + elog(ERROR,"Arguments array are not the same type!"); + + if (ARRISVOID(a) || ARRISVOID(b)) + PG_RETURN_FLOAT4(0.0); + + info = sa->info; + if (info->tupDesc == NULL) + elog(ERROR, "Only weigthed (composite) types should be used"); + getFmgrInfoCmp(info); + + while(ai < sa->nelems && bi < sb->nelems) + { + Datum ad = deconstructCompositeType(info, sa->elems[ai], &tmpA), + bd = deconstructCompositeType(info, sb->elems[bi], &tmpB); + + cmp = DatumGetInt32(FCall2(&info->cmpFunc, ad, bd)); + + if ( cmp < 0 ) { + if (useIntersect == false) + denominatorA += tmpA * tmpA; + ai++; + } else if ( cmp > 0 ) { + if (useIntersect == false) + denominatorB += tmpB * tmpB; + bi++; + } else { + denominatorA += tmpA * tmpA; + denominatorB += tmpB * tmpB; + numerator += tmpA * tmpB; + ai++; + bi++; + } + } + + if (useIntersect == false) { + while(ai < sa->nelems) { + deconstructCompositeType(info, sa->elems[ai], &tmpA); + denominatorA += tmpA * tmpA; + ai++; + } + + while(bi < sb->nelems) { + deconstructCompositeType(info, sb->elems[bi], &tmpB); + denominatorB += tmpB * tmpB; + bi++; + } + } + + if (numerator != 0.0) { + numerator = numerator / sqrt( denominatorA * denominatorB ); + } + + PG_RETURN_FLOAT4(numerator); +} + +PG_FUNCTION_INFO_V1(arraysml_op); +Datum arraysml_op(PG_FUNCTION_ARGS); +Datum +arraysml_op(PG_FUNCTION_ARGS) +{ + ArrayType *a, *b; + SimpleArray *sa, *sb; + double power = 0.0; + + fcinfo->flinfo->fn_extra = SearchArrayCache( + fcinfo->flinfo->fn_extra, + fcinfo->flinfo->fn_mcxt, + PG_GETARG_DATUM(0), &a, &sa, NULL); + fcinfo->flinfo->fn_extra = SearchArrayCache( + fcinfo->flinfo->fn_extra, + fcinfo->flinfo->fn_mcxt, + PG_GETARG_DATUM(1), &b, &sb, NULL); + + if ( ARR_ELEMTYPE(a) != ARR_ELEMTYPE(b) ) + elog(ERROR,"Arguments array are not the same type!"); + + if (ARRISVOID(a) || ARRISVOID(b)) + PG_RETURN_BOOL(false); + + switch(getSmlType()) + { + case ST_TFIDF: + power = TFIDFSml(sa, sb); + break; + case ST_COSINE: + { + int cnt; + + power = sqrt( ((double)(sa->nelems)) * ((double)(sb->nelems)) ); + + if ( ((double)Min(sa->nelems, sb->nelems)) / power < GetSmlarLimit() ) + PG_RETURN_BOOL(false); + + cnt = numOfIntersect(sa, sb); + power = ((double)cnt) / power; + } + break; + case ST_OVERLAP: + power = (double)numOfIntersect(sa, sb); + break; + default: + elog(ERROR,"Unsupported formula type of similarity"); + } + + PG_RETURN_BOOL(power >= GetSmlarLimit()); +} + +#define QBSIZE 8192 +static char cachedFormula[QBSIZE]; +static int cachedLen = 0; +static void *cachedPlan = NULL; + +PG_FUNCTION_INFO_V1(arraysml_func); +Datum arraysml_func(PG_FUNCTION_ARGS); +Datum +arraysml_func(PG_FUNCTION_ARGS) +{ + ArrayType *a, *b; + SimpleArray *sa, *sb; + int cnt; + float4 result = -1.0; + Oid arg[] = {INT4OID, INT4OID, INT4OID}; + Datum pars[3]; + bool isnull; + void *plan; + int stat; + text *formula = PG_GETARG_TEXT_P(2); + + fcinfo->flinfo->fn_extra = SearchArrayCache( + fcinfo->flinfo->fn_extra, + fcinfo->flinfo->fn_mcxt, + PG_GETARG_DATUM(0), &a, &sa, NULL); + fcinfo->flinfo->fn_extra = SearchArrayCache( + fcinfo->flinfo->fn_extra, + fcinfo->flinfo->fn_mcxt, + PG_GETARG_DATUM(1), &b, &sb, NULL); + + if ( ARR_ELEMTYPE(a) != ARR_ELEMTYPE(b) ) + elog(ERROR,"Arguments array are not the same type!"); + + if (ARRISVOID(a) || ARRISVOID(b)) + PG_RETURN_BOOL(false); + + cnt = numOfIntersect(sa, sb); + + if ( VARSIZE(formula) - VARHDRSZ > QBSIZE - 1024 ) + elog(ERROR,"Formula is too long"); + + SPI_connect(); + + if ( cachedPlan == NULL || cachedLen != VARSIZE(formula) - VARHDRSZ || + memcmp( cachedFormula, VARDATA(formula), VARSIZE(formula) - VARHDRSZ ) != 0 ) + { + char *ptr, buf[QBSIZE]; + + *cachedFormula = '\0'; + if ( cachedPlan ) + SPI_freeplan(cachedPlan); + cachedPlan = NULL; + cachedLen = 0; + + ptr = stpcpy( buf, "SELECT (" ); + memcpy( ptr, VARDATA(formula), VARSIZE(formula) - VARHDRSZ ); + ptr += VARSIZE(formula) - VARHDRSZ; + ptr = stpcpy( ptr, ")::float4 FROM"); + ptr = stpcpy( ptr, " (SELECT $1 ::float8 AS i, $2 ::float8 AS a, $3 ::float8 AS b) AS N;"); + *ptr = '\0'; + + plan = SPI_prepare(buf, 3, arg); + if (!plan) + elog(ERROR, "SPI_prepare() failed"); + + cachedPlan = SPI_saveplan(plan); + if (!cachedPlan) + elog(ERROR, "SPI_saveplan() failed"); + + SPI_freeplan(plan); + cachedLen = VARSIZE(formula) - VARHDRSZ; + memcpy( cachedFormula, VARDATA(formula), VARSIZE(formula) - VARHDRSZ ); + } + + plan = cachedPlan; + + + pars[0] = Int32GetDatum( cnt ); + pars[1] = Int32GetDatum( sa->nelems ); + pars[2] = Int32GetDatum( sb->nelems ); + + stat = SPI_execute_plan(plan, pars, NULL, true, 3); + if (stat < 0) + elog(ERROR, "SPI_execute_plan() returns %d", stat); + + if ( SPI_processed > 0) + result = DatumGetFloat4(SPI_getbinval(SPI_tuptable->vals[0], SPI_tuptable->tupdesc, 1, &isnull)); + + SPI_finish(); + + PG_RETURN_FLOAT4(result); +} + +PG_FUNCTION_INFO_V1(array_unique); +Datum array_unique(PG_FUNCTION_ARGS); +Datum +array_unique(PG_FUNCTION_ARGS) +{ + ArrayType *a = PG_GETARG_ARRAYTYPE_P(0); + ArrayType *res; + SimpleArray *sa; + + sa = Array2SimpleArrayU(NULL, a, NULL); + + res = construct_array( sa->elems, + sa->nelems, + sa->info->typid, + sa->info->typlen, + sa->info->typbyval, + sa->info->typalign); + + pfree(sa->elems); + pfree(sa); + PG_FREE_IF_COPY(a, 0); + + PG_RETURN_ARRAYTYPE_P(res); +} + +PG_FUNCTION_INFO_V1(inarray); +Datum inarray(PG_FUNCTION_ARGS); +Datum +inarray(PG_FUNCTION_ARGS) +{ + ArrayType *a; + SimpleArray *sa; + Datum query = PG_GETARG_DATUM(1); + Oid queryTypeOid; + Datum *StopLow, + *StopHigh, + *StopMiddle; + int cmp; + + fcinfo->flinfo->fn_extra = SearchArrayCache( + fcinfo->flinfo->fn_extra, + fcinfo->flinfo->fn_mcxt, + PG_GETARG_DATUM(0), &a, &sa, NULL); + + queryTypeOid = get_fn_expr_argtype(fcinfo->flinfo, 1); + + if ( queryTypeOid == InvalidOid ) + elog(ERROR,"inarray: could not determine actual argument type"); + + if ( queryTypeOid != sa->info->typid ) + elog(ERROR,"inarray: Type of array's element and type of argument are not the same"); + + getFmgrInfoCmp(sa->info); + StopLow = sa->elems; + StopHigh = sa->elems + sa->nelems; + + while (StopLow < StopHigh) + { + StopMiddle = StopLow + ((StopHigh - StopLow) >> 1); + cmp = cmpArrayElem(StopMiddle, &query, sa->info); + + if ( cmp == 0 ) + { + /* found */ + if ( PG_NARGS() >= 3 ) + PG_RETURN_DATUM(PG_GETARG_DATUM(2)); + PG_RETURN_FLOAT4(1.0); + } + else if (cmp < 0) + StopLow = StopMiddle + 1; + else + StopHigh = StopMiddle; + } + + if ( PG_NARGS() >= 4 ) + PG_RETURN_DATUM(PG_GETARG_DATUM(3)); + PG_RETURN_FLOAT4(0.0); +} diff --git a/external/smlar/smlar.control b/external/smlar/smlar.control new file mode 100644 index 00000000000..e6aa6e97b27 --- /dev/null +++ b/external/smlar/smlar.control @@ -0,0 +1,5 @@ +comment = 'compute similary of any one-dimensional arrays' +default_version = '1.0' +module_pathname = '$libdir/smlar' +relocatable = true + diff --git a/external/smlar/smlar.h b/external/smlar/smlar.h new file mode 100644 index 00000000000..6410326c3d7 --- /dev/null +++ b/external/smlar/smlar.h @@ -0,0 +1,144 @@ +#ifndef _SMLAR_H_ +#define _SMLAR_H_ + +#include "postgres.h" +#include "utils/array.h" +#include "access/tupdesc.h" +#include "catalog/pg_collation.h" + +#include + +typedef struct ProcTypeInfoData *ProcTypeInfo; + +typedef struct ProcTypeInfoData { + Oid typid; + Oid hashFuncOid; + Oid cmpFuncOid; + int16 typlen; + bool typbyval; + char typalign; + + /* support of composite type */ + char typtype; + TupleDesc tupDesc; + + /* + * Following member can become invalid, + * so fill it just before using + */ + bool hashFuncInited; + FmgrInfo hashFunc; + bool cmpFuncInited; + FmgrInfo cmpFunc; +} ProcTypeInfoData; + +ProcTypeInfo findProcs(Oid typid); +void getFmgrInfoHash(ProcTypeInfo info); +void getFmgrInfoCmp(ProcTypeInfo info); + +#define NDIM 1 +/* reject arrays we can't handle; but allow a NULL or empty array */ +#define CHECKARRVALID(x) \ + do { \ + if (x) { \ + if (ARR_NDIM(x) != NDIM && ARR_NDIM(x) != 0) \ + ereport(ERROR, \ + (errcode(ERRCODE_ARRAY_SUBSCRIPT_ERROR), \ + errmsg("array must be one-dimensional"))); \ + if (ARR_HASNULL(x)) \ + ereport(ERROR, \ + (errcode(ERRCODE_NULL_VALUE_NOT_ALLOWED), \ + errmsg("array must not contain nulls"))); \ + } \ + } while(0) + +#define ARRISVOID(x) ((x) == NULL || ARRNELEMS(x) == 0) +#define ARRNELEMS(x) ArrayGetNItems(ARR_NDIM(x), ARR_DIMS(x)) + + +typedef struct SimpleArray { + Datum *elems; + double *df; /* frequency in current doc */ + uint32 *hash; + int nelems; + ProcTypeInfo info; +} SimpleArray; + +SimpleArray * Array2SimpleArray(ProcTypeInfo info, ArrayType *a); +SimpleArray * Array2SimpleArrayS(ProcTypeInfo info, ArrayType *a); +SimpleArray * Array2SimpleArrayU(ProcTypeInfo info, ArrayType *a, void *cache); +void allocateHash(void *cache, SimpleArray *a); + +/* + * GUC vars + */ +double GetSmlarLimit(void); +const char* GetSmlarTable(void); +bool GetSmlarUsePersistent(void); +double getOneAdd(void); +int getTFMethod(void); +int getSmlType(void); +/* + * GiST + */ + +#define SmlarOverlapStrategy 1 +#define SmlarSimilarityStrategy 2 + +struct SmlSign; +struct SmlSign* Array2HashedArray(ProcTypeInfo info, ArrayType *a); +/* + * Cache subsystem + */ +void* SearchArrayCache( void *cache, MemoryContext ctx, Datum a, ArrayType **da, SimpleArray **sa, struct SmlSign **ss ); + +typedef struct StatElem { + Datum datum; + double idf; /* log(d/df) */ +} StatElem; + +typedef struct HashedElem { + uint32 hash; + double idfMin; + double idfMax; +} HashedElem; + +typedef struct SignedElem { + double idfMin; + double idfMax; +} SignedElem; + +typedef struct StatCache { + StatElem *elems; + int nelems; + int64_t ndoc; + HashedElem *helems; + int nhelems; + SignedElem *selems; + ProcTypeInfo info; +} StatCache; + +StatCache *initStatCache(MemoryContext ctx); +void getHashStatCache(StatCache *stat, MemoryContext ctx, size_t n); + +void resetStatCache(void); +StatElem *findStat(StatCache *stat, Datum query, StatElem *low); +StatElem *fingArrayStat(void *cache, Oid typoid, Datum query, StatElem *low); +StatCache *getStat(void *cache, size_t n); + +/* + * Formula's type of similarity + */ +#define ST_COSINE 1 +#define ST_TFIDF 2 +#define ST_OVERLAP 3 +/* + * TF methods + */ +#define TF_N 1 +#define TF_LOG 2 +#define TF_CONST 3 + +#define FCall2(f, x1, x2) FunctionCall2Coll((f), C_COLLATION_OID, (x1), (x2)) + +#endif diff --git a/external/smlar/smlar.sql.in.in b/external/smlar/smlar.sql.in.in new file mode 100644 index 00000000000..636a989ca13 --- /dev/null +++ b/external/smlar/smlar.sql.in.in @@ -0,0 +1,713 @@ +SET search_path = public; + +CREATE OR REPLACE FUNCTION smlar(anyarray, anyarray) + RETURNS float4 + AS 'MODULE_PATHNAME', 'arraysml' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION smlar(anyarray, anyarray, bool) + RETURNS float4 + AS 'MODULE_PATHNAME', 'arraysmlw' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION smlar(anyarray, anyarray, text) + RETURNS float4 + AS 'MODULE_PATHNAME', 'arraysml_func' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION set_smlar_limit(float4) + RETURNS float4 + AS 'MODULE_PATHNAME' + LANGUAGE C STRICT VOLATILE; + +CREATE OR REPLACE FUNCTION show_smlar_limit() + RETURNS float4 + AS 'MODULE_PATHNAME' + LANGUAGE C STRICT STABLE; + +CREATE OR REPLACE FUNCTION smlar_op(anyarray,anyarray) + RETURNS bool + AS 'MODULE_PATHNAME', 'arraysml_op' + LANGUAGE C STRICT STABLE; + +CREATE OPERATOR % ( + LEFTARG = anyarray, + RIGHTARG = anyarray, + PROCEDURE = smlar_op, + COMMUTATOR = '%', + RESTRICT = contsel, + JOIN = contjoinsel +); + +-- + +CREATE OR REPLACE FUNCTION tsvector2textarray(tsvector) + RETURNS _text + AS 'MODULE_PATHNAME', 'tsvector2textarray' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION array_unique(anyarray) + RETURNS anyarray + AS 'MODULE_PATHNAME', 'array_unique' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION inarray(anyarray, anyelement) + RETURNS float4 + AS 'MODULE_PATHNAME', 'inarray' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION inarray(anyarray, anyelement, float4, float4) + RETURNS float4 + AS 'MODULE_PATHNAME', 'inarray' + LANGUAGE C STRICT IMMUTABLE; + +--gist key + +CREATE OR REPLACE FUNCTION gsmlsign_in(cstring) + RETURNS gsmlsign + AS 'MODULE_PATHNAME' + LANGUAGE C STRICT; + +CREATE OR REPLACE FUNCTION gsmlsign_out(gsmlsign) + RETURNS cstring + AS 'MODULE_PATHNAME' + LANGUAGE C STRICT; + +CREATE TYPE gsmlsign ( + INTERNALLENGTH = -1, + INPUT = gsmlsign_in, + OUTPUT = gsmlsign_out +); + +-- support functions for gist +CREATE OR REPLACE FUNCTION gsmlsign_consistent(gsmlsign,internal,int4) + RETURNS bool + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_compress(internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_decompress(internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_penalty(internal,internal,internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C STRICT IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_picksplit(internal, internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_union(bytea, internal) + RETURNS _int4 + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION gsmlsign_same(gsmlsign, gsmlsign, internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +--create the operator classes for gist + +CREATE OPERATOR CLASS _int2_sml_ops +FOR TYPE _int2 USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _int4_sml_ops +FOR TYPE _int4 USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _int8_sml_ops +FOR TYPE _int8 USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _float4_sml_ops +FOR TYPE _float4 USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _float8_sml_ops +FOR TYPE _float8 USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +-- money type hasn't hash support +--CREATE OPERATOR CLASS _money_sml_ops +--FOR TYPE _money USING gist +--AS +-- OPERATOR 1 && (anyarray, anyarray), +-- OPERATOR 2 % (anyarray, anyarray), +-- FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), +-- FUNCTION 2 gsmlsign_union (bytea, internal), +-- FUNCTION 3 gsmlsign_compress (internal), +-- FUNCTION 4 gsmlsign_decompress (internal), +-- FUNCTION 5 gsmlsign_penalty (internal, internal, internal), +-- FUNCTION 6 gsmlsign_picksplit (internal, internal), +-- FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +--STORAGE gsmlsign; + +CREATE OPERATOR CLASS _oid_sml_ops +FOR TYPE _oid USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _timestamp_sml_ops +FOR TYPE _timestamp USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _timestamptz_sml_ops +FOR TYPE _timestamptz USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _time_sml_ops +FOR TYPE _time USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _timetz_sml_ops +FOR TYPE _timetz USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _date_sml_ops +FOR TYPE _date USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _interval_sml_ops +FOR TYPE _interval USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _macaddr_sml_ops +FOR TYPE _macaddr USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _inet_sml_ops +FOR TYPE _inet USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _cidr_sml_ops +FOR TYPE _cidr USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _text_sml_ops +FOR TYPE _text USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _varchar_sml_ops +FOR TYPE _varchar USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _char_sml_ops +FOR TYPE _char USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +CREATE OPERATOR CLASS _bytea_sml_ops +FOR TYPE _bytea USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +--no hash +--CREATE OPERATOR CLASS _bit_sml_ops +--FOR TYPE _bit USING gist +--AS +-- OPERATOR 1 && (anyarray, anyarray), +-- OPERATOR 2 % (anyarray, anyarray), +-- FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), +-- FUNCTION 2 gsmlsign_union (bytea, internal), +-- FUNCTION 3 gsmlsign_compress (internal), +-- FUNCTION 4 gsmlsign_decompress (internal), +-- FUNCTION 5 gsmlsign_penalty (internal, internal, internal), +-- FUNCTION 6 gsmlsign_picksplit (internal, internal), +-- FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +--STORAGE gsmlsign; + +--CREATE OPERATOR CLASS _varbit_sml_ops +--FOR TYPE _varbit USING gist +--AS +-- OPERATOR 1 && (anyarray, anyarray), +-- OPERATOR 2 % (anyarray, anyarray), +-- FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), +-- FUNCTION 2 gsmlsign_union (bytea, internal), +-- FUNCTION 3 gsmlsign_compress (internal), +-- FUNCTION 4 gsmlsign_decompress (internal), +-- FUNCTION 5 gsmlsign_penalty (internal, internal, internal), +-- FUNCTION 6 gsmlsign_picksplit (internal, internal), +-- FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +--STORAGE gsmlsign; + +CREATE OPERATOR CLASS _numeric_sml_ops +FOR TYPE _numeric USING gist +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 gsmlsign_consistent (gsmlsign, internal, int4), + FUNCTION 2 gsmlsign_union (bytea, internal), + FUNCTION 3 gsmlsign_compress (internal), + FUNCTION 4 gsmlsign_decompress (internal), + FUNCTION 5 gsmlsign_penalty (internal, internal, internal), + FUNCTION 6 gsmlsign_picksplit (internal, internal), + FUNCTION 7 gsmlsign_same (gsmlsign, gsmlsign, internal), +STORAGE gsmlsign; + +--gin support functions +CREATE OR REPLACE FUNCTION smlararrayextract(anyarray, internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION smlarqueryarrayextract(anyarray, internal, internal) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +CREATE OR REPLACE FUNCTION smlararrayconsistent(internal, internal, anyarray) + RETURNS internal + AS 'MODULE_PATHNAME' + LANGUAGE C IMMUTABLE; + +--gin opclasses + +CREATE OPERATOR CLASS _int2_sml_ops +FOR TYPE _int2 USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btint2cmp(int2,int2), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE int2; +CREATE OPERATOR CLASS _int4_sml_ops +FOR TYPE _int4 USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btint4cmp(int4,int4), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE int4; + +CREATE OPERATOR CLASS _int8_sml_ops +FOR TYPE _int8 USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btint8cmp(int8,int8), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE int8; + +CREATE OPERATOR CLASS _float4_sml_ops +FOR TYPE _float4 USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btfloat4cmp(float4,float4), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE float4; + +CREATE OPERATOR CLASS _float8_sml_ops +FOR TYPE _float8 USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btfloat8cmp(float8,float8), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE float8; + +CREATE OPERATOR CLASS _money_sml_ops +FOR TYPE _money USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 cash_cmp(money,money), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE money; + +CREATE OPERATOR CLASS _oid_sml_ops +FOR TYPE _oid USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btoidcmp(oid,oid), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE oid; + +CREATE OPERATOR CLASS _timestamp_sml_ops +FOR TYPE _timestamp USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 timestamp_cmp(timestamp,timestamp), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE timestamp; + +CREATE OPERATOR CLASS _timestamptz_sml_ops +FOR TYPE _timestamptz USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 timestamptz_cmp(timestamptz,timestamptz), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE timestamptz; + +CREATE OPERATOR CLASS _time_sml_ops +FOR TYPE _time USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 time_cmp(time,time), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE time; + +CREATE OPERATOR CLASS _timetz_sml_ops +FOR TYPE _timetz USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 timetz_cmp(timetz,timetz), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE timetz; + +CREATE OPERATOR CLASS _date_sml_ops +FOR TYPE _date USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 date_cmp(date,date), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE date; + +CREATE OPERATOR CLASS _interval_sml_ops +FOR TYPE _interval USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 interval_cmp(interval,interval), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE interval; + +CREATE OPERATOR CLASS _macaddr_sml_ops +FOR TYPE _macaddr USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 macaddr_cmp(macaddr,macaddr), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE macaddr; + +CREATE OPERATOR CLASS _inet_sml_ops +FOR TYPE _inet USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 network_cmp(inet,inet), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE inet; + +CREATE OPERATOR CLASS _cidr_sml_ops +FOR TYPE _cidr USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 network_cmp(inet,inet), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE cidr; + +CREATE OPERATOR CLASS _text_sml_ops +FOR TYPE _text USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 bttextcmp(text,text), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE text; + +CREATE OPERATOR CLASS _varchar_sml_ops +FOR TYPE _varchar USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 bttextcmp(text,text), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE varchar; + +CREATE OPERATOR CLASS _char_sml_ops +FOR TYPE "_char" USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 btcharcmp("char","char"), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE "char"; + +CREATE OPERATOR CLASS _bytea_sml_ops +FOR TYPE _bytea USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 byteacmp(bytea,bytea), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE bytea; + +CREATE OPERATOR CLASS _bit_sml_ops +FOR TYPE _bit USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 bitcmp(bit,bit), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE bit; + +CREATE OPERATOR CLASS _varbit_sml_ops +FOR TYPE _varbit USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 varbitcmp(varbit,varbit), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE varbit; + +CREATE OPERATOR CLASS _numeric_sml_ops +FOR TYPE _numeric USING gin +AS + OPERATOR 1 && (anyarray, anyarray), + OPERATOR 2 % (anyarray, anyarray), + FUNCTION 1 numeric_cmp(numeric,numeric), + FUNCTION 2 smlararrayextract(anyarray, internal), + FUNCTION 3 smlarqueryarrayextract(anyarray, internal, internal), + FUNCTION 4 smlararrayconsistent(internal, internal, anyarray), + STORAGE numeric; + diff --git a/external/smlar/smlar_cache.c b/external/smlar/smlar_cache.c new file mode 100644 index 00000000000..edd593dd426 --- /dev/null +++ b/external/smlar/smlar_cache.c @@ -0,0 +1,301 @@ +#include "smlar.h" + +#include "fmgr.h" +#include "utils/array.h" +#include "utils/memutils.h" + +/* + * Deconstructed array cache + */ + +typedef struct ArrayCacheEntry { + Datum toastedArray; + ArrayType *da; + SimpleArray *sa; + struct SmlSign *ss; + + struct ArrayCacheEntry *prev; + struct ArrayCacheEntry *next; +} ArrayCacheEntry; + +#define NENTRIES (16) + +typedef struct ArrayCache { + MemoryContext ctx; + int32 nentries; + ArrayCacheEntry *head; + ArrayCacheEntry *tail; + ArrayCacheEntry* entries[ NENTRIES ]; + StatCache *DocStat; +} ArrayCache; + +static void +moveFirst(ArrayCache *ac, ArrayCacheEntry *entry) +{ + /* + * delete entry form a list + */ + Assert( entry != ac->head ); + if ( entry == ac->tail ) + { + Assert( entry->next == NULL ); + ac->tail = entry->prev; + if ( ac->tail ) + ac->tail->next = NULL; + else + ac->head = NULL; + } + else + { + entry->prev->next = entry->next; + entry->next->prev = entry->prev; + } + + /* + * Install into head + */ + + Assert( ac->head != NULL ); + Assert( ac->tail != NULL ); + + entry->next = ac->head; + entry->prev = NULL; + ac->head->prev = entry; + ac->head = entry; +} + +#define DATUMSIZE(d) VARSIZE_ANY(DatumGetPointer(d)) +static int +cmpDatum(Datum a, Datum b) +{ + int32 la = DATUMSIZE(a); + int32 lb = DATUMSIZE(b); + + if ( la == lb ) + return memcmp( DatumGetPointer(a), DatumGetPointer(b), la ); + + return (la > lb) ? 1 : -1; +} + +static void +fetchData(ArrayCache *ac, ArrayCacheEntry *entry, ArrayType **da, SimpleArray **sa, struct SmlSign **ss ) +{ + ProcTypeInfo info; + + info = findProcs( ARR_ELEMTYPE(entry->da) ); + + if ( sa ) + { + if ( entry->sa == NULL ) + { + MemoryContext old; + + getFmgrInfoCmp(info); + + old = MemoryContextSwitchTo( ac->ctx ); + entry->sa = Array2SimpleArrayU(info, entry->da, (getSmlType() == ST_TFIDF) ? ac : NULL); + MemoryContextSwitchTo(old); + } + + *sa = entry->sa; + } + + if ( ss ) + { + if ( entry->ss == NULL ) + { + MemoryContext old; + + getFmgrInfoHash(info); + + old = MemoryContextSwitchTo( ac->ctx ); + entry->ss = Array2HashedArray(info, entry->da); + MemoryContextSwitchTo(old); + } + + *ss = entry->ss; + } + + if (da) + *da = entry->da; +} + +static void +cleanupData(ArrayCacheEntry *entry) +{ + if ( entry->sa ) + { + if ( entry->sa->elems ) + pfree( entry->sa->elems ); + if ( entry->sa->df ) + pfree( entry->sa->df ); + if ( entry->sa->hash ) + pfree( entry->sa->hash ); + pfree( entry->sa ); + } + entry->sa = NULL; + + if ( entry->ss ) + pfree(entry->ss); + entry->ss = NULL; + + pfree( DatumGetPointer(entry->toastedArray) ); + pfree( entry->da ); +} + +static void +makeEntry(ArrayCache *ac, ArrayCacheEntry *entry, Datum a) +{ + ArrayType *detoastedArray; + + entry->toastedArray = (Datum)MemoryContextAlloc( ac->ctx, DATUMSIZE(a) ); + memcpy( DatumGetPointer(entry->toastedArray), DatumGetPointer(a), DATUMSIZE(a) ); + + detoastedArray = (ArrayType*)PG_DETOAST_DATUM(entry->toastedArray); + entry->da = MemoryContextAlloc( ac->ctx, VARSIZE(detoastedArray)); + memcpy( entry->da, detoastedArray, VARSIZE(detoastedArray)); +} + +static int +cmpEntry(const void *a, const void *b) +{ + return cmpDatum( (*(ArrayCacheEntry**)a)->toastedArray, (*(ArrayCacheEntry**)b)->toastedArray ); +} + +void* +SearchArrayCache( void *cache, MemoryContext ctx, Datum a, ArrayType **da, SimpleArray **sa, struct SmlSign **ss ) +{ + ArrayCache *ac; + ArrayCacheEntry *entry; + + if ( cache == NULL ) + cache = MemoryContextAllocZero(ctx, sizeof(ArrayCache)); + + ac = (ArrayCache*) cache; + ac->ctx = ctx; + + /* + * Fast check of resent used value + */ + if ( ac->head && cmpDatum(ac->head->toastedArray, a) == 0 ) + { + fetchData(ac, ac->head, da, sa, ss); + return cache; + } + + if ( ac->head == NULL ) + { + ac->entries[0] = ac->head = ac->tail = MemoryContextAllocZero(ctx, sizeof(ArrayCacheEntry)); + ac->nentries = 1; + makeEntry(ac, ac->head, a); + fetchData(ac, ac->head, da, sa, ss); + return cache; + } + + do { + ArrayCacheEntry **StopLow = ac->entries; + ArrayCacheEntry **StopHigh = ac->entries + ac->nentries; + ArrayCacheEntry **StopMiddle; + int cmp; + + while (StopLow < StopHigh) { + StopMiddle = StopLow + ((StopHigh - StopLow) >> 1); + entry = *StopMiddle; + cmp = cmpDatum(entry->toastedArray, a); + + if ( cmp == 0 ) + { + moveFirst(ac, entry); + fetchData(ac, ac->head, da, sa, ss); + return cache; + } + else if ( cmp < 0 ) + StopLow = StopMiddle + 1; + else + StopHigh = StopMiddle; + } + } while(0); + + /* + * Not found + */ + + if ( ac->nentries < NENTRIES ) + { + entry = ac->entries[ ac->nentries ] = MemoryContextAllocZero(ctx, sizeof(ArrayCacheEntry)); + + /* install first */ + entry->next = ac->head; + entry->prev = NULL; + ac->head->prev = entry; + ac->head = entry; + + ac->nentries ++; + + makeEntry(ac, ac->head, a); + fetchData(ac, ac->head, da, sa, ss); + } + else + { + cleanupData( ac->tail ); + moveFirst(ac, ac->tail ); + makeEntry(ac, ac->head, a); + fetchData(ac, ac->head, da, sa, ss); + } + + qsort(ac->entries, ac->nentries, sizeof(ArrayCacheEntry*), cmpEntry); + return cache; +} + +StatElem * +fingArrayStat(void *cache, Oid typoid, Datum query, StatElem *low) +{ + ArrayCache *ac; + + if ( cache == NULL ) + return NULL; + + ac = (ArrayCache*) cache; + if ( ac->DocStat == NULL ) + { + ac->DocStat = initStatCache(ac->ctx); + low = NULL; + } + + if ( typoid != ac->DocStat->info->typid ) + elog(ERROR,"Types of stat table and actual arguments are different"); + + return findStat(ac->DocStat, query, low); +} + +StatCache * +getStat(void *cache, size_t n) +{ + ArrayCache *ac; + + if ( cache == NULL ) + return NULL; + + ac = (ArrayCache*) cache; + if ( ac->DocStat == NULL ) + ac->DocStat = initStatCache(ac->ctx); + + getHashStatCache(ac->DocStat, ac->ctx, n); + + return ac->DocStat; +} + +void +allocateHash(void *cache, SimpleArray *a) +{ + ArrayCache *ac; + + if ( cache == NULL ) + return; + ac = (ArrayCache*) cache; + + if (a->hash) + return; + + a->hash = MemoryContextAlloc( ac->ctx, sizeof(uint32) * a->nelems ); +} diff --git a/external/smlar/smlar_gin.c b/external/smlar/smlar_gin.c new file mode 100644 index 00000000000..4164fa4bfdc --- /dev/null +++ b/external/smlar/smlar_gin.c @@ -0,0 +1,152 @@ +#include + +#include "smlar.h" + +#include "fmgr.h" +#include "access/gin.h" +#include "access/skey.h" +#if PG_VERSION_NUM < 130000 +#include "access/tuptoaster.h" +#else +#include "access/heaptoast.h" +#endif + +PG_FUNCTION_INFO_V1(smlararrayextract); +Datum smlararrayextract(PG_FUNCTION_ARGS); +Datum +smlararrayextract(PG_FUNCTION_ARGS) +{ + ArrayType *array; + int32 *nentries = (int32 *) PG_GETARG_POINTER(1); + SimpleArray *sa; + + /* + * we should guarantee that array will not be destroyed during all + * operation + */ + array = PG_GETARG_ARRAYTYPE_P_COPY(0); + + CHECKARRVALID(array); + + sa = Array2SimpleArrayU(NULL, array, NULL); + + *nentries = sa->nelems; + + if (sa->nelems == 0 && PG_NARGS() == 3) + { + switch (PG_GETARG_UINT16(2)) /* StrategyNumber */ + { + case SmlarOverlapStrategy: + case SmlarSimilarityStrategy: + *nentries = -1; /* nobody can be found */ + break; + default: + break; + } + } + + PG_RETURN_POINTER( sa->elems ); +} + +PG_FUNCTION_INFO_V1(smlarqueryarrayextract); +Datum smlarqueryarrayextract(PG_FUNCTION_ARGS); +Datum +smlarqueryarrayextract(PG_FUNCTION_ARGS) +{ + PG_RETURN_DATUM(DirectFunctionCall3(smlararrayextract, + PG_GETARG_DATUM(0), + PG_GETARG_DATUM(1), + PG_GETARG_DATUM(2))); +} + +PG_FUNCTION_INFO_V1(smlararrayconsistent); +Datum smlararrayconsistent(PG_FUNCTION_ARGS); +Datum +smlararrayconsistent(PG_FUNCTION_ARGS) +{ + bool *check = (bool *) PG_GETARG_POINTER(0); + StrategyNumber strategy = PG_GETARG_UINT16(1); + SimpleArray *sa; + bool res = false; + int i, + cnt = 0; + bool *recheck = (bool *) PG_GETARG_POINTER(5); + + *recheck = true; + + switch (strategy) + { + case SmlarOverlapStrategy: + /* at least one element in check[] is true, so result = true */ + res = true; + *recheck = false; + break; + case SmlarSimilarityStrategy: + + fcinfo->flinfo->fn_extra = SearchArrayCache( + fcinfo->flinfo->fn_extra, + fcinfo->flinfo->fn_mcxt, + PG_GETARG_DATUM(2), NULL, &sa, NULL ); + + for(i=0; inelems; i++) + cnt += check[i]; + + /* + * cnt is a lower limit of elements's number in indexed array; + */ + + switch(getSmlType()) + { + case ST_TFIDF: + { + double weight = 0.0, /* exact weight of union */ + saSum = 0.0, /* exact length of query */ + siSum = 0.0; /* lower limit of length of indexed value */ + + if ( getTFMethod() != TF_CONST ) + elog(ERROR,"GIN supports only smlar.tf_method = \"const\"" ); + + Assert(sa->df); + + for(i=0; inelems; i++) + { + /* + * With smlar.tf_method = "const" sa->df[i] is + * equal to its idf, so lookup of StatElem is not needed + */ + if ( check[i] ) + { + weight += sa->df[i] * sa->df[i]; + siSum += sa->df[i] * sa->df[i]; + } + saSum += sa->df[i] * sa->df[i]; + } + + if ( saSum > 0.0 && siSum > 0.0 && weight / sqrt(saSum * siSum ) > GetSmlarLimit() ) + res = true; + } + break; + case ST_COSINE: + { + double power; + + power = sqrt( ((double)(sa->nelems)) * ((double)(cnt)) ); + + if ( ((double)cnt) / power >= GetSmlarLimit() ) + res = true; + } + break; + case ST_OVERLAP: + if (cnt >= GetSmlarLimit()) + res = true; + break; + default: + elog(ERROR,"GIN doesn't support current formula type of similarity"); + } + break; + default: + elog(ERROR, "smlararrayconsistent: unknown strategy number: %d", strategy); + } + + PG_RETURN_BOOL(res); +} diff --git a/external/smlar/smlar_gist.c b/external/smlar/smlar_gist.c new file mode 100644 index 00000000000..480d74451c7 --- /dev/null +++ b/external/smlar/smlar_gist.c @@ -0,0 +1,1251 @@ +#include "smlar.h" + +#include "fmgr.h" +#include "access/gist.h" +#include "access/skey.h" +#if PG_VERSION_NUM < 130000 +#include "access/tuptoaster.h" +#else +#include "access/heaptoast.h" +#endif +#include "utils/memutils.h" + +typedef struct SmlSign { + int32 vl_len_; /* varlena header (do not touch directly!) */ + int32 flag:8, + size:24; + int32 maxrepeat; + char data[1]; +} SmlSign; + +#define SMLSIGNHDRSZ (offsetof(SmlSign, data)) + +#define BITBYTE 8 +#define SIGLENINT 61 +#define SIGLEN ( sizeof(int)*SIGLENINT ) +#define SIGLENBIT (SIGLEN*BITBYTE - 1) /* see makesign */ +typedef char BITVEC[SIGLEN]; +typedef char *BITVECP; +#define LOOPBYTE \ + for(i=0;i> i & 0x01 ) +#define CLRBIT(x,i) GETBYTE(x,i) &= ~( 0x01 << ( (i) % BITBYTE ) ) +#define SETBIT(x,i) GETBYTE(x,i) |= ( 0x01 << ( (i) % BITBYTE ) ) +#define GETBIT(x,i) ( (GETBYTE(x,i) >> ( (i) % BITBYTE )) & 0x01 ) + +#define HASHVAL(val) (((unsigned int)(val)) % SIGLENBIT) +#define HASH(sign, val) SETBIT((sign), HASHVAL(val)) + +#define ARRKEY 0x01 +#define SIGNKEY 0x02 +#define ALLISTRUE 0x04 + +#define ISARRKEY(x) ( ((SmlSign*)x)->flag & ARRKEY ) +#define ISSIGNKEY(x) ( ((SmlSign*)x)->flag & SIGNKEY ) +#define ISALLTRUE(x) ( ((SmlSign*)x)->flag & ALLISTRUE ) + +#define CALCGTSIZE(flag, len) ( SMLSIGNHDRSZ + ( ( (flag) & ARRKEY ) ? ((len)*sizeof(uint32)) : (((flag) & ALLISTRUE) ? 0 : SIGLEN) ) ) +#define GETSIGN(x) ( (BITVECP)( (char*)x+SMLSIGNHDRSZ ) ) +#define GETARR(x) ( (uint32*)( (char*)x+SMLSIGNHDRSZ ) ) + +#define GETENTRY(vec,pos) ((SmlSign *) DatumGetPointer((vec)->vector[(pos)].key)) + +/* + * Fake IO + */ +PG_FUNCTION_INFO_V1(gsmlsign_in); +Datum gsmlsign_in(PG_FUNCTION_ARGS); +Datum +gsmlsign_in(PG_FUNCTION_ARGS) +{ + elog(ERROR, "not implemented"); + PG_RETURN_DATUM(0); +} + +PG_FUNCTION_INFO_V1(gsmlsign_out); +Datum gsmlsign_out(PG_FUNCTION_ARGS); +Datum +gsmlsign_out(PG_FUNCTION_ARGS) +{ + elog(ERROR, "not implemented"); + PG_RETURN_DATUM(0); +} + +/* + * Compress/decompress + */ + +/* Number of one-bits in an unsigned byte */ +static const uint8 number_of_ones[256] = { + 0, 1, 1, 2, 1, 2, 2, 3, 1, 2, 2, 3, 2, 3, 3, 4, + 1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5, + 1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5, + 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, + 1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5, + 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, + 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, + 3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7, + 1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5, + 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, + 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, + 3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7, + 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, + 3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7, + 3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7, + 4, 5, 5, 6, 5, 6, 6, 7, 5, 6, 6, 7, 6, 7, 7, 8 +}; + +static int +compareint(const void *va, const void *vb) +{ + uint32 a = *((uint32 *) va); + uint32 b = *((uint32 *) vb); + + if (a == b) + return 0; + return (a > b) ? 1 : -1; +} + +/* + * Removes duplicates from an array of int32. 'l' is + * size of the input array. Returns the new size of the array. + */ +static int +uniqueint(uint32 *a, int32 l, int32 *max) +{ + uint32 *ptr, + *res; + int32 cnt = 0; + + *max = 1; + + if (l <= 1) + return l; + + ptr = res = a; + + qsort((void *) a, l, sizeof(uint32), compareint); + + while (ptr - a < l) + if (*ptr != *res) + { + cnt = 1; + *(++res) = *ptr++; + } + else + { + cnt++; + if ( cnt > *max ) + *max = cnt; + ptr++; + } + + if ( cnt > *max ) + *max = cnt; + + return res + 1 - a; +} + +SmlSign* +Array2HashedArray(ProcTypeInfo info, ArrayType *a) +{ + SimpleArray *s = Array2SimpleArray(info, a); + SmlSign *sign; + int32 len, i; + uint32 *ptr; + + len = CALCGTSIZE( ARRKEY, s->nelems ); + + getFmgrInfoHash(s->info); + if (s->info->tupDesc) + elog(ERROR, "GiST doesn't support composite (weighted) type"); + + sign = palloc( len ); + sign->flag = ARRKEY; + sign->size = s->nelems; + + ptr = GETARR(sign); + for(i=0;inelems;i++) + ptr[i] = DatumGetUInt32(FunctionCall1Coll(&s->info->hashFunc, + DEFAULT_COLLATION_OID, + s->elems[i])); + + /* + * there is a collision of hash-function; len is always equal or less than + * s->nelems + */ + sign->size = uniqueint( GETARR(sign), sign->size, &sign->maxrepeat ); + len = CALCGTSIZE( ARRKEY, sign->size ); + SET_VARSIZE(sign, len); + + return sign; +} + +static int +HashedElemCmp(const void *va, const void *vb) +{ + uint32 a = ((HashedElem *) va)->hash; + uint32 b = ((HashedElem *) vb)->hash; + + if (a == b) + { + double ma = ((HashedElem *) va)->idfMin; + double mb = ((HashedElem *) va)->idfMin; + + if (ma == mb) + return 0; + + return ( ma > mb ) ? 1 : -1; + } + + return (a > b) ? 1 : -1; +} + +static int +uniqueHashedElem(HashedElem *a, int32 l) +{ + HashedElem *ptr, + *res; + + if (l <= 1) + return l; + + ptr = res = a; + + qsort(a, l, sizeof(HashedElem), HashedElemCmp); + + while (ptr - a < l) + if (ptr->hash != res->hash) + *(++res) = *ptr++; + else + { + res->idfMax = ptr->idfMax; + ptr++; + } + + return res + 1 - a; +} + +static StatCache* +getHashedCache(void *cache) +{ + StatCache *stat = getStat(cache, SIGLENBIT); + + if ( stat->nhelems < 0 ) + { + int i; + /* + * Init + */ + + if (stat->info->tupDesc) + elog(ERROR, "GiST doesn't support composite (weighted) type"); + getFmgrInfoHash(stat->info); + for(i=0;inelems;i++) + { + uint32 hash; + int index; + + hash = DatumGetUInt32(FunctionCall1Coll(&stat->info->hashFunc, + DEFAULT_COLLATION_OID, + stat->elems[i].datum)); + index = HASHVAL(hash); + + stat->helems[i].hash = hash; + stat->helems[i].idfMin = stat->helems[i].idfMax = stat->elems[i].idf; + if ( stat->selems[index].idfMin == 0.0 ) + stat->selems[index].idfMin = stat->selems[index].idfMax = stat->elems[i].idf; + else if ( stat->selems[index].idfMin > stat->elems[i].idf ) + stat->selems[index].idfMin = stat->elems[i].idf; + else if ( stat->selems[index].idfMax < stat->elems[i].idf ) + stat->selems[index].idfMax = stat->elems[i].idf; + } + + stat->nhelems = uniqueHashedElem( stat->helems, stat->nelems); + } + + return stat; +} + +static HashedElem* +getHashedElemIdf(StatCache *stat, uint32 hash, HashedElem *StopLow) +{ + HashedElem *StopMiddle, + *StopHigh = stat->helems + stat->nhelems; + + if ( !StopLow ) + StopLow = stat->helems; + + while (StopLow < StopHigh) { + StopMiddle = StopLow + ((StopHigh - StopLow) >> 1); + + if ( StopMiddle->hash == hash ) + return StopMiddle; + else if ( StopMiddle->hash < hash ) + StopLow = StopMiddle + 1; + else + StopHigh = StopMiddle; + } + + return NULL; +} + +static void +fillHashVal(void *cache, SimpleArray *a) +{ + int i; + + if (a->hash) + return; + + allocateHash(cache, a); + + if (a->info->tupDesc) + elog(ERROR, "GiST doesn't support composite (weighted) type"); + getFmgrInfoHash(a->info); + + for(i=0;inelems;i++) + a->hash[i] = DatumGetUInt32(FunctionCall1Coll(&a->info->hashFunc, + DEFAULT_COLLATION_OID, + a->elems[i])); +} + + +static bool +hasHashedElem(SmlSign *a, uint32 h) +{ + uint32 *StopLow = GETARR(a), + *StopHigh = GETARR(a) + a->size, + *StopMiddle; + + while (StopLow < StopHigh) { + StopMiddle = StopLow + ((StopHigh - StopLow) >> 1); + + if ( *StopMiddle == h ) + return true; + else if ( *StopMiddle < h ) + StopLow = StopMiddle + 1; + else + StopHigh = StopMiddle; + } + + return false; +} + +static void +makesign(BITVECP sign, SmlSign *a) +{ + int32 i; + uint32 *ptr = GETARR(a); + + MemSet((void *) sign, 0, sizeof(BITVEC)); + SETBIT(sign, SIGLENBIT); /* set last unused bit */ + + for (i = 0; i < a->size; i++) + HASH(sign, ptr[i]); +} + +static int32 +sizebitvec(BITVECP sign) +{ + int32 size = 0, + i; + + LOOPBYTE + size += number_of_ones[(unsigned char) sign[i]]; + + return size; +} + +PG_FUNCTION_INFO_V1(gsmlsign_compress); +Datum gsmlsign_compress(PG_FUNCTION_ARGS); +Datum +gsmlsign_compress(PG_FUNCTION_ARGS) +{ + GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); + GISTENTRY *retval = entry; + + if (entry->leafkey) /* new key */ + { + SmlSign *sign; + ArrayType *a = DatumGetArrayTypeP(entry->key); + + sign = Array2HashedArray(NULL, a); + + if ( VARSIZE(sign) > TOAST_INDEX_TARGET ) + { /* make signature due to its big size */ + SmlSign *tmpsign; + int len; + + len = CALCGTSIZE( SIGNKEY, sign->size ); + tmpsign = palloc( len ); + tmpsign->flag = SIGNKEY; + SET_VARSIZE(tmpsign, len); + + makesign(GETSIGN(tmpsign), sign); + tmpsign->size = sizebitvec(GETSIGN(tmpsign)); + tmpsign->maxrepeat = sign->maxrepeat; + sign = tmpsign; + } + + retval = (GISTENTRY *) palloc(sizeof(GISTENTRY)); + gistentryinit(*retval, PointerGetDatum(sign), + entry->rel, entry->page, + entry->offset, false); + } + else if ( ISSIGNKEY(DatumGetPointer(entry->key)) && + !ISALLTRUE(DatumGetPointer(entry->key)) ) + { + SmlSign *sign = (SmlSign*)DatumGetPointer(entry->key); + + Assert( sign->size == sizebitvec(GETSIGN(sign)) ); + + if ( sign->size == SIGLENBIT ) + { + int32 len = CALCGTSIZE(SIGNKEY | ALLISTRUE, 0); + int32 maxrepeat = sign->maxrepeat; + + sign = (SmlSign *) palloc(len); + SET_VARSIZE(sign, len); + sign->flag = SIGNKEY | ALLISTRUE; + sign->size = SIGLENBIT; + sign->maxrepeat = maxrepeat; + + retval = (GISTENTRY *) palloc(sizeof(GISTENTRY)); + + gistentryinit(*retval, PointerGetDatum(sign), + entry->rel, entry->page, + entry->offset, false); + } + } + + PG_RETURN_POINTER(retval); +} + +PG_FUNCTION_INFO_V1(gsmlsign_decompress); +Datum gsmlsign_decompress(PG_FUNCTION_ARGS); +Datum +gsmlsign_decompress(PG_FUNCTION_ARGS) +{ + GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); + SmlSign *key = (SmlSign*)DatumGetPointer(PG_DETOAST_DATUM(entry->key)); + + if (key != (SmlSign *) DatumGetPointer(entry->key)) + { + GISTENTRY *retval = (GISTENTRY *) palloc(sizeof(GISTENTRY)); + + gistentryinit(*retval, PointerGetDatum(key), + entry->rel, entry->page, + entry->offset, false); + + PG_RETURN_POINTER(retval); + } + + PG_RETURN_POINTER(entry); +} + +/* + * Union method + */ +static bool +unionkey(BITVECP sbase, SmlSign *add) +{ + int32 i; + + if (ISSIGNKEY(add)) + { + BITVECP sadd = GETSIGN(add); + + if (ISALLTRUE(add)) + return true; + + LOOPBYTE + sbase[i] |= sadd[i]; + } + else + { + uint32 *ptr = GETARR(add); + + for (i = 0; i < add->size; i++) + HASH(sbase, ptr[i]); + } + + return false; +} + +PG_FUNCTION_INFO_V1(gsmlsign_union); +Datum gsmlsign_union(PG_FUNCTION_ARGS); +Datum +gsmlsign_union(PG_FUNCTION_ARGS) +{ + GistEntryVector *entryvec = (GistEntryVector *) PG_GETARG_POINTER(0); + int *size = (int *) PG_GETARG_POINTER(1); + BITVEC base; + int32 i, + len, + maxrepeat = 1; + int32 flag = 0; + SmlSign *result; + + MemSet((void *) base, 0, sizeof(BITVEC)); + for (i = 0; i < entryvec->n; i++) + { + if (GETENTRY(entryvec, i)->maxrepeat > maxrepeat) + maxrepeat = GETENTRY(entryvec, i)->maxrepeat; + if (unionkey(base, GETENTRY(entryvec, i))) + { + flag = ALLISTRUE; + break; + } + } + + flag |= SIGNKEY; + len = CALCGTSIZE(flag, 0); + result = (SmlSign *) palloc(len); + *size = len; + SET_VARSIZE(result, len); + result->flag = flag; + result->maxrepeat = maxrepeat; + + if (!ISALLTRUE(result)) + { + memcpy((void *) GETSIGN(result), (void *) base, sizeof(BITVEC)); + result->size = sizebitvec(GETSIGN(result)); + } + else + result->size = SIGLENBIT; + + PG_RETURN_POINTER(result); +} + +/* + * Same method + */ + +PG_FUNCTION_INFO_V1(gsmlsign_same); +Datum gsmlsign_same(PG_FUNCTION_ARGS); +Datum +gsmlsign_same(PG_FUNCTION_ARGS) +{ + SmlSign *a = (SmlSign*)PG_GETARG_POINTER(0); + SmlSign *b = (SmlSign*)PG_GETARG_POINTER(1); + bool *result = (bool *) PG_GETARG_POINTER(2); + + if (a->size != b->size) + { + *result = false; + } + else if (ISSIGNKEY(a)) + { /* then b also ISSIGNKEY */ + if ( ISALLTRUE(a) ) + { + /* in this case b is all true too - other cases is catched + upper */ + *result = true; + } + else + { + int32 i; + BITVECP sa = GETSIGN(a), + sb = GETSIGN(b); + + *result = true; + + if ( !ISALLTRUE(a) ) + { + LOOPBYTE + { + if (sa[i] != sb[i]) + { + *result = false; + break; + } + } + } + } + } + else + { + uint32 *ptra = GETARR(a), + *ptrb = GETARR(b); + int32 i; + + *result = true; + for (i = 0; i < a->size; i++) + { + if ( ptra[i] != ptrb[i]) + { + *result = false; + break; + } + } + } + + PG_RETURN_POINTER(result); +} + +/* + * Penalty method + */ +static int +hemdistsign(BITVECP a, BITVECP b) +{ + int i, + diff, + dist = 0; + + LOOPBYTE + { + diff = (unsigned char) (a[i] ^ b[i]); + dist += number_of_ones[diff]; + } + return dist; +} + +static int +hemdist(SmlSign *a, SmlSign *b) +{ + if (ISALLTRUE(a)) + { + if (ISALLTRUE(b)) + return 0; + else + return SIGLENBIT - b->size; + } + else if (ISALLTRUE(b)) + return SIGLENBIT - b->size; + + return hemdistsign(GETSIGN(a), GETSIGN(b)); +} + +PG_FUNCTION_INFO_V1(gsmlsign_penalty); +Datum gsmlsign_penalty(PG_FUNCTION_ARGS); +Datum +gsmlsign_penalty(PG_FUNCTION_ARGS) +{ + GISTENTRY *origentry = (GISTENTRY *) PG_GETARG_POINTER(0); /* always ISSIGNKEY */ + GISTENTRY *newentry = (GISTENTRY *) PG_GETARG_POINTER(1); + float *penalty = (float *) PG_GETARG_POINTER(2); + SmlSign *origval = (SmlSign *) DatumGetPointer(origentry->key); + SmlSign *newval = (SmlSign *) DatumGetPointer(newentry->key); + BITVECP orig = GETSIGN(origval); + + *penalty = 0.0; + + if (ISARRKEY(newval)) + { + BITVEC sign; + + makesign(sign, newval); + + if (ISALLTRUE(origval)) + *penalty = ((float) (SIGLENBIT - sizebitvec(sign))) / (float) (SIGLENBIT + 1); + else + *penalty = hemdistsign(sign, orig); + } + else + *penalty = hemdist(origval, newval); + + PG_RETURN_POINTER(penalty); +} + +/* + * Picksplit method + */ + +typedef struct +{ + bool allistrue; + int32 size; + BITVEC sign; +} CACHESIGN; + +static void +fillcache(CACHESIGN *item, SmlSign *key) +{ + item->allistrue = false; + item->size = key->size; + + if (ISARRKEY(key)) + { + makesign(item->sign, key); + item->size = sizebitvec( item->sign ); + } + else if (ISALLTRUE(key)) + item->allistrue = true; + else + memcpy((void *) item->sign, (void *) GETSIGN(key), sizeof(BITVEC)); +} + +#define WISH_F(a,b,c) (double)( -(double)(((a)-(b))*((a)-(b))*((a)-(b)))*(c) ) + +typedef struct +{ + OffsetNumber pos; + int32 cost; +} SPLITCOST; + +static int +comparecost(const void *va, const void *vb) +{ + SPLITCOST *a = (SPLITCOST *) va; + SPLITCOST *b = (SPLITCOST *) vb; + + if (a->cost == b->cost) + return 0; + else + return (a->cost > b->cost) ? 1 : -1; +} + +static int +hemdistcache(CACHESIGN *a, CACHESIGN *b) +{ + if (a->allistrue) + { + if (b->allistrue) + return 0; + else + return SIGLENBIT - b->size; + } + else if (b->allistrue) + return SIGLENBIT - a->size; + + return hemdistsign(a->sign, b->sign); +} + +PG_FUNCTION_INFO_V1(gsmlsign_picksplit); +Datum gsmlsign_picksplit(PG_FUNCTION_ARGS); +Datum +gsmlsign_picksplit(PG_FUNCTION_ARGS) +{ + GistEntryVector *entryvec = (GistEntryVector *) PG_GETARG_POINTER(0); + GIST_SPLITVEC *v = (GIST_SPLITVEC *) PG_GETARG_POINTER(1); + OffsetNumber k, + j; + SmlSign *datum_l, + *datum_r; + BITVECP union_l, + union_r; + int32 size_alpha, + size_beta; + int32 size_waste, + waste = -1; + int32 nbytes; + OffsetNumber seed_1 = 0, + seed_2 = 0; + OffsetNumber *left, + *right; + OffsetNumber maxoff; + BITVECP ptr; + int i; + CACHESIGN *cache; + SPLITCOST *costvector; + + maxoff = entryvec->n - 2; + nbytes = (maxoff + 2) * sizeof(OffsetNumber); + v->spl_left = (OffsetNumber *) palloc(nbytes); + v->spl_right = (OffsetNumber *) palloc(nbytes); + + cache = (CACHESIGN *) palloc(sizeof(CACHESIGN) * (maxoff + 2)); + fillcache(&cache[FirstOffsetNumber], GETENTRY(entryvec, FirstOffsetNumber)); + + for (k = FirstOffsetNumber; k < maxoff; k = OffsetNumberNext(k)) + { + for (j = OffsetNumberNext(k); j <= maxoff; j = OffsetNumberNext(j)) + { + if (k == FirstOffsetNumber) + fillcache(&cache[j], GETENTRY(entryvec, j)); + + size_waste = hemdistcache(&(cache[j]), &(cache[k])); + if (size_waste > waste) + { + waste = size_waste; + seed_1 = k; + seed_2 = j; + } + } + } + + left = v->spl_left; + v->spl_nleft = 0; + right = v->spl_right; + v->spl_nright = 0; + + if (seed_1 == 0 || seed_2 == 0) + { + seed_1 = 1; + seed_2 = 2; + } + + /* form initial .. */ + if (cache[seed_1].allistrue) + { + datum_l = (SmlSign *) palloc(CALCGTSIZE(SIGNKEY | ALLISTRUE, 0)); + SET_VARSIZE(datum_l, CALCGTSIZE(SIGNKEY | ALLISTRUE, 0)); + datum_l->flag = SIGNKEY | ALLISTRUE; + datum_l->size = SIGLENBIT; + } + else + { + datum_l = (SmlSign *) palloc(CALCGTSIZE(SIGNKEY, 0)); + SET_VARSIZE(datum_l, CALCGTSIZE(SIGNKEY, 0)); + datum_l->flag = SIGNKEY; + memcpy((void *) GETSIGN(datum_l), (void *) cache[seed_1].sign, sizeof(BITVEC)); + datum_l->size = cache[seed_1].size; + } + if (cache[seed_2].allistrue) + { + datum_r = (SmlSign *) palloc(CALCGTSIZE(SIGNKEY | ALLISTRUE, 0)); + SET_VARSIZE(datum_r, CALCGTSIZE(SIGNKEY | ALLISTRUE, 0)); + datum_r->flag = SIGNKEY | ALLISTRUE; + datum_r->size = SIGLENBIT; + } + else + { + datum_r = (SmlSign *) palloc(CALCGTSIZE(SIGNKEY, 0)); + SET_VARSIZE(datum_r, CALCGTSIZE(SIGNKEY, 0)); + datum_r->flag = SIGNKEY; + memcpy((void *) GETSIGN(datum_r), (void *) cache[seed_2].sign, sizeof(BITVEC)); + datum_r->size = cache[seed_2].size; + } + + union_l = GETSIGN(datum_l); + union_r = GETSIGN(datum_r); + maxoff = OffsetNumberNext(maxoff); + fillcache(&cache[maxoff], GETENTRY(entryvec, maxoff)); + /* sort before ... */ + costvector = (SPLITCOST *) palloc(sizeof(SPLITCOST) * maxoff); + for (j = FirstOffsetNumber; j <= maxoff; j = OffsetNumberNext(j)) + { + costvector[j - 1].pos = j; + size_alpha = hemdistcache(&(cache[seed_1]), &(cache[j])); + size_beta = hemdistcache(&(cache[seed_2]), &(cache[j])); + costvector[j - 1].cost = Abs(size_alpha - size_beta); + } + qsort((void *) costvector, maxoff, sizeof(SPLITCOST), comparecost); + + datum_l->maxrepeat = datum_r->maxrepeat = 1; + + for (k = 0; k < maxoff; k++) + { + j = costvector[k].pos; + if (j == seed_1) + { + *left++ = j; + v->spl_nleft++; + continue; + } + else if (j == seed_2) + { + *right++ = j; + v->spl_nright++; + continue; + } + + if (ISALLTRUE(datum_l) || cache[j].allistrue) + { + if (ISALLTRUE(datum_l) && cache[j].allistrue) + size_alpha = 0; + else + size_alpha = SIGLENBIT - ( + (cache[j].allistrue) ? datum_l->size : cache[j].size + ); + } + else + size_alpha = hemdistsign(cache[j].sign, GETSIGN(datum_l)); + + if (ISALLTRUE(datum_r) || cache[j].allistrue) + { + if (ISALLTRUE(datum_r) && cache[j].allistrue) + size_beta = 0; + else + size_beta = SIGLENBIT - ( + (cache[j].allistrue) ? datum_r->size : cache[j].size + ); + } + else + size_beta = hemdistsign(cache[j].sign, GETSIGN(datum_r)); + + if (size_alpha < size_beta + WISH_F(v->spl_nleft, v->spl_nright, 0.1)) + { + if (ISALLTRUE(datum_l) || cache[j].allistrue) + { + if (!ISALLTRUE(datum_l)) + MemSet((void *) GETSIGN(datum_l), 0xff, sizeof(BITVEC)); + datum_l->size = SIGLENBIT; + } + else + { + ptr = cache[j].sign; + LOOPBYTE + union_l[i] |= ptr[i]; + datum_l->size = sizebitvec(union_l); + } + *left++ = j; + v->spl_nleft++; + } + else + { + if (ISALLTRUE(datum_r) || cache[j].allistrue) + { + if (!ISALLTRUE(datum_r)) + MemSet((void *) GETSIGN(datum_r), 0xff, sizeof(BITVEC)); + datum_r->size = SIGLENBIT; + } + else + { + ptr = cache[j].sign; + LOOPBYTE + union_r[i] |= ptr[i]; + datum_r->size = sizebitvec(union_r); + } + *right++ = j; + v->spl_nright++; + } + } + *right = *left = FirstOffsetNumber; + v->spl_ldatum = PointerGetDatum(datum_l); + v->spl_rdatum = PointerGetDatum(datum_r); + + Assert( datum_l->size = sizebitvec(GETSIGN(datum_l)) ); + Assert( datum_r->size = sizebitvec(GETSIGN(datum_r)) ); + + PG_RETURN_POINTER(v); +} + +static double +getIdfMaxLimit(SmlSign *key) +{ + switch( getTFMethod() ) + { + case TF_CONST: + return 1.0; + break; + case TF_N: + return (double)(key->maxrepeat); + break; + case TF_LOG: + return 1.0 + log( (double)(key->maxrepeat) ); + break; + default: + elog(ERROR,"Unknown TF method: %d", getTFMethod()); + } + + return 0.0; +} + +/* + * Consistent function + */ +PG_FUNCTION_INFO_V1(gsmlsign_consistent); +Datum gsmlsign_consistent(PG_FUNCTION_ARGS); +Datum +gsmlsign_consistent(PG_FUNCTION_ARGS) +{ + GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0); + StrategyNumber strategy = (StrategyNumber) PG_GETARG_UINT16(2); + bool *recheck = (bool *) PG_GETARG_POINTER(4); + ArrayType *a; + SmlSign *key = (SmlSign*)DatumGetPointer(entry->key); + int res = false; + SmlSign *query; + SimpleArray *s; + int32 i; + + fcinfo->flinfo->fn_extra = SearchArrayCache( + fcinfo->flinfo->fn_extra, + fcinfo->flinfo->fn_mcxt, + PG_GETARG_DATUM(1), &a, &s, &query); + + *recheck = true; + + if ( ARRISVOID(a) ) + PG_RETURN_BOOL(res); + + if ( strategy == SmlarOverlapStrategy ) + { + if (ISALLTRUE(key)) + { + res = true; + } + else if (ISARRKEY(key)) + { + uint32 *kptr = GETARR(key), + *qptr = GETARR(query); + + while( kptr - GETARR(key) < key->size && qptr - GETARR(query) < query->size ) + { + if ( *kptr < *qptr ) + kptr++; + else if ( *kptr > *qptr ) + qptr++; + else + { + res = true; + break; + } + } + *recheck = false; + } + else + { + BITVECP sign = GETSIGN(key); + + fillHashVal(fcinfo->flinfo->fn_extra, s); + + for(i=0; inelems; i++) + { + if ( GETBIT(sign, HASHVAL(s->hash[i])) ) + { + res = true; + break; + } + } + } + } + /* + * SmlarSimilarityStrategy + */ + else if (ISALLTRUE(key)) + { + if ( GIST_LEAF(entry) ) + { + /* + * With TF/IDF similarity we cannot say anything useful + */ + if ( query->size < SIGLENBIT && getSmlType() != ST_TFIDF ) + { + double power = ((double)(query->size)) * ((double)(SIGLENBIT)); + + if ( ((double)(query->size)) / sqrt(power) >= GetSmlarLimit() ) + res = true; + } + else + { + res = true; + } + } + else + res = true; + } + else if (ISARRKEY(key)) + { + uint32 *kptr = GETARR(key), + *qptr = GETARR(query); + + Assert( GIST_LEAF(entry) ); + + switch(getSmlType()) + { + case ST_TFIDF: + { + StatCache *stat = getHashedCache(fcinfo->flinfo->fn_extra); + double sumU = 0.0, + sumQ = 0.0, + sumK = 0.0; + double maxKTF = getIdfMaxLimit(key); + HashedElem *h; + + Assert( s->df ); + fillHashVal(fcinfo->flinfo->fn_extra, s); + if ( stat->info != s->info ) + elog(ERROR,"Statistic and actual argument have different type"); + + for(i=0;inelems;i++) + { + sumQ += s->df[i] * s->df[i]; + + h = getHashedElemIdf(stat, s->hash[i], NULL); + if ( h && hasHashedElem(key, s->hash[i]) ) + { + sumK += h->idfMin * h->idfMin; + sumU += h->idfMax * maxKTF * s->df[i]; + } + } + + if ( sumK > 0.0 && sumQ > 0.0 && sumU / sqrt( sumK * sumQ ) >= GetSmlarLimit() ) + { + /* + * More precisely calculate sumK + */ + h = NULL; + sumK = 0.0; + + for(i=0;isize;i++) + { + h = getHashedElemIdf(stat, GETARR(key)[i], h); + if (h) + sumK += h->idfMin * h->idfMin; + } + + if ( sumK > 0.0 && sumQ > 0.0 && sumU / sqrt( sumK * sumQ ) >= GetSmlarLimit() ) + res = true; + } + } + break; + case ST_COSINE: + { + double power; + power = sqrt( ((double)(key->size)) * ((double)(s->nelems)) ); + + if ( ((double)Min(key->size, s->nelems)) / power >= GetSmlarLimit() ) + { + int cnt = 0; + + while( kptr - GETARR(key) < key->size && qptr - GETARR(query) < query->size ) + { + if ( *kptr < *qptr ) + kptr++; + else if ( *kptr > *qptr ) + qptr++; + else + { + cnt++; + kptr++; + qptr++; + } + } + + if ( ((double)cnt) / power >= GetSmlarLimit() ) + res = true; + } + } + break; + default: + elog(ERROR,"GiST doesn't support current formula type of similarity"); + } + } + else + { /* signature */ + BITVECP sign = GETSIGN(key); + int32 count = 0; + + fillHashVal(fcinfo->flinfo->fn_extra, s); + + if ( GIST_LEAF(entry) ) + { + switch(getSmlType()) + { + case ST_TFIDF: + { + StatCache *stat = getHashedCache(fcinfo->flinfo->fn_extra); + double sumU = 0.0, + sumQ = 0.0, + sumK = 0.0; + double maxKTF = getIdfMaxLimit(key); + + Assert( s->df ); + if ( stat->info != s->info ) + elog(ERROR,"Statistic and actual argument have different type"); + + for(i=0;inelems;i++) + { + int32 hbit = HASHVAL(s->hash[i]); + + sumQ += s->df[i] * s->df[i]; + if ( GETBIT(sign, hbit) ) + { + sumK += stat->selems[ hbit ].idfMin * stat->selems[ hbit ].idfMin; + sumU += stat->selems[ hbit ].idfMax * maxKTF * s->df[i]; + } + } + + if ( sumK > 0.0 && sumQ > 0.0 && sumU / sqrt( sumK * sumQ ) >= GetSmlarLimit() ) + { + /* + * More precisely calculate sumK + */ + sumK = 0.0; + + for(i=0;iselems[ i ].idfMin * stat->selems[ i ].idfMin; + + if ( sumK > 0.0 && sumQ > 0.0 && sumU / sqrt( sumK * sumQ ) >= GetSmlarLimit() ) + res = true; + } + } + break; + case ST_COSINE: + { + double power; + + power = sqrt( ((double)(key->size)) * ((double)(s->nelems)) ); + + for(i=0; inelems; i++) + count += GETBIT(sign, HASHVAL(s->hash[i])); + + if ( ((double)count) / power >= GetSmlarLimit() ) + res = true; + } + break; + default: + elog(ERROR,"GiST doesn't support current formula type of similarity"); + } + } + else /* non-leaf page */ + { + switch(getSmlType()) + { + case ST_TFIDF: + { + StatCache *stat = getHashedCache(fcinfo->flinfo->fn_extra); + double sumU = 0.0, + sumQ = 0.0, + minK = -1.0; + double maxKTF = getIdfMaxLimit(key); + + Assert( s->df ); + if ( stat->info != s->info ) + elog(ERROR,"Statistic and actual argument have different type"); + + for(i=0;inelems;i++) + { + int32 hbit = HASHVAL(s->hash[i]); + + sumQ += s->df[i] * s->df[i]; + if ( GETBIT(sign, hbit) ) + { + sumU += stat->selems[ hbit ].idfMax * maxKTF * s->df[i]; + if ( minK > stat->selems[ hbit ].idfMin || minK < 0.0 ) + minK = stat->selems[ hbit ].idfMin; + } + } + + if ( sumQ > 0.0 && minK > 0.0 && sumU / sqrt( sumQ * minK ) >= GetSmlarLimit() ) + res = true; + } + break; + case ST_COSINE: + { + for(i=0; inelems; i++) + count += GETBIT(sign, HASHVAL(s->hash[i])); + + if ( s->nelems == count || sqrt(((double)count) / ((double)(s->nelems))) >= GetSmlarLimit() ) + res = true; + } + break; + default: + elog(ERROR,"GiST doesn't support current formula type of similarity"); + } + } + } + +#if 0 + { + static int nnres = 0; + static int nres = 0; + if ( GIST_LEAF(entry) ) { + if ( res ) + nres++; + else + nnres++; + elog(NOTICE,"%s nn:%d n:%d", (ISARRKEY(key)) ? "ARR" : ( (ISALLTRUE(key)) ? "TRUE" : "SIGN" ), nnres, nres ); + } + } +#endif + + PG_RETURN_BOOL(res); +} diff --git a/external/smlar/smlar_guc.c b/external/smlar/smlar_guc.c new file mode 100644 index 00000000000..8ff0ad3d675 --- /dev/null +++ b/external/smlar/smlar_guc.c @@ -0,0 +1,229 @@ +#include "smlar.h" + +#include "fmgr.h" +#include "utils/guc.h" + +#include "smlar.h" + +/* + * Smlar limit + */ +static double smlar_limit = 0.6; + +/* + * Smlar table stores table-wide statistic + */ +static char * smlar_table = ""; + +static bool smlar_use_malloc = false; + +/* + * State of GUC initialization + */ +static bool smlar_guc_inited = false; + +static void +SmlarTableAssign(const char *newval, void *extra) +{ + resetStatCache(); +} + +static bool smlar_logadd = false; + +static void +SmlarLogAssign(bool newval, void *extra) +{ + resetStatCache(); +} + +static int smlar_smltype = ST_COSINE; + +static const struct config_enum_entry SmlarTypeOptions[] = { + {"cosine", ST_COSINE, false}, + {"tfidf", ST_TFIDF, false}, + {"overlap", ST_OVERLAP, false}, + {NULL, 0, false} +}; + +static int smlar_tf_method = TF_N; +static const struct config_enum_entry SmlarTFOptions[] = { + {"n", TF_N, false}, + {"log", TF_LOG, false}, + {"const", TF_CONST, false}, + {NULL, 0, false} +}; + +static void +initSmlarGUC() +{ + if (smlar_guc_inited) + return; + + DefineCustomRealVariable( + "smlar.threshold", + "Lower threshold of array's similarity", + "Array's with similarity lower than threshold are not similar by % operation", + &smlar_limit, + 0.6, + 0.0, + 1e10, + PGC_USERSET, + 0, + NULL, + NULL, + NULL + ); + + DefineCustomStringVariable( + "smlar.stattable", + "Name of table stored set-wide statistic", + "Named table stores global frequencies of array's elements", + &smlar_table, + "", + PGC_USERSET, + GUC_IS_NAME, + NULL, + SmlarTableAssign, + NULL + ); + + DefineCustomEnumVariable( + "smlar.type", + "Type of similarity formula", + "Type of similarity formula: cosine(default), tfidf, overlap", + &smlar_smltype, + smlar_smltype, + SmlarTypeOptions, + PGC_SUSET, + 0, + NULL, + NULL, + NULL + ); + + DefineCustomBoolVariable( + "smlar.persistent_cache", + "Usage of persistent cache of global stat", + "Cache of global stat is stored in transaction-independent memory", + &smlar_use_malloc, + false, + PGC_USERSET, + 0, + NULL, + NULL, + NULL + ); + + DefineCustomBoolVariable( + "smlar.idf_plus_one", + "Calculate idf by log(1+d/df)", + "Calculate idf by log(1+d/df)", + &smlar_logadd, + false, + PGC_USERSET, + 0, + NULL, + SmlarLogAssign, + NULL + ); + + DefineCustomEnumVariable( + "smlar.tf_method", + "Method of TF caclulation", + "TF method: n => number of entries, log => 1+log(n), const => constant value", + &smlar_tf_method, + smlar_tf_method, + SmlarTFOptions, + PGC_SUSET, + 0, + NULL, + NULL, + NULL + ); + + smlar_guc_inited = true; +} + +double +getOneAdd(void) +{ + if (!smlar_guc_inited) + initSmlarGUC(); + + return (smlar_logadd) ? 1.0 : 0.0; +} + +int +getTFMethod(void) +{ + if (!smlar_guc_inited) + initSmlarGUC(); + + return smlar_tf_method; +} + +double +GetSmlarLimit(void) +{ + if (!smlar_guc_inited) + initSmlarGUC(); + + return smlar_limit; +} + +const char* +GetSmlarTable(void) +{ + if (!smlar_guc_inited) + initSmlarGUC(); + + return smlar_table; +} + +int +getSmlType(void) +{ + if (!smlar_guc_inited) + initSmlarGUC(); + + return smlar_smltype; +} + +bool +GetSmlarUsePersistent(void) +{ + if (!smlar_guc_inited) + initSmlarGUC(); + + return smlar_use_malloc; +} + +PG_FUNCTION_INFO_V1(set_smlar_limit); +Datum set_smlar_limit(PG_FUNCTION_ARGS); +Datum +set_smlar_limit(PG_FUNCTION_ARGS) +{ + float4 nlimit = PG_GETARG_FLOAT4(0); + char buf[32]; + + /* init smlar guc */ + initSmlarGUC(); + + sprintf(buf,"%f", nlimit); + set_config_option("smlar.threshold", buf, + PGC_USERSET, PGC_S_SESSION ,GUC_ACTION_SET, true, 0 +#if PG_VERSION_NUM >= 90500 + ,false +#endif + ); + PG_RETURN_FLOAT4((float4)GetSmlarLimit()); +} + +PG_FUNCTION_INFO_V1(show_smlar_limit); +Datum show_smlar_limit(PG_FUNCTION_ARGS); +Datum +show_smlar_limit(PG_FUNCTION_ARGS) +{ + PG_RETURN_FLOAT4((float4)GetSmlarLimit()); +} + diff --git a/external/smlar/smlar_stat.c b/external/smlar/smlar_stat.c new file mode 100644 index 00000000000..9a563aaca3d --- /dev/null +++ b/external/smlar/smlar_stat.c @@ -0,0 +1,206 @@ +#include "smlar.h" + +#include "fmgr.h" +#include "catalog/pg_type.h" +#include "executor/spi.h" +#include "utils/array.h" +#include "utils/datum.h" +#include "utils/memutils.h" + +static StatCache *PersistentDocStat = NULL; + +static void* +cacheAlloc(MemoryContext ctx, size_t size) +{ + if ( GetSmlarUsePersistent() ) + { + void *ptr = malloc(size); + + if (!ptr) + ereport(ERROR, + (errcode(ERRCODE_OUT_OF_MEMORY), + errmsg("out of memory"))); + + return ptr; + } + + return MemoryContextAlloc(ctx, size); +} + +static void* +cacheAllocZero(MemoryContext ctx, size_t size) +{ + void *ptr = cacheAlloc(ctx, size); + memset(ptr, 0, size); + return ptr; +} + +struct StatCache * +initStatCache(MemoryContext ctx) +{ + if (PersistentDocStat && GetSmlarUsePersistent()) + return PersistentDocStat; + else { + int stat; + char buf[1024]; + const char *tbl = GetSmlarTable(); + StatCache *cache = NULL; + + if ( tbl == NULL || *tbl == '\0' ) + elog(ERROR,"smlar.stattable is not defined"); + + sprintf(buf,"SELECT * FROM \"%s\" ORDER BY 1;", tbl); + SPI_connect(); + stat = SPI_execute(buf, true, 0); + + if (stat != SPI_OK_SELECT) + elog(ERROR, "SPI_execute() returns %d", stat); + + if ( SPI_processed == 0 ) + { + elog(ERROR, "Stat table '%s' is empty", tbl); + } + else + { + int i; + double totaldocs = 0.0; + Oid ndocType = SPI_gettypeid(SPI_tuptable->tupdesc, 2); + + if ( SPI_tuptable->tupdesc->natts != 2 ) + elog(ERROR,"Stat table is not (type, int4)"); + if ( !(ndocType == INT4OID || ndocType == INT8OID) ) + elog(ERROR,"Stat table is not (type, int4) nor (type, int8)"); + + cache = cacheAllocZero(ctx, sizeof(StatCache)); + cache->info = findProcs( SPI_gettypeid(SPI_tuptable->tupdesc, 1) ); + if (cache->info->tupDesc) + elog(ERROR, "TF/IDF is not supported for composite (weighted) type"); + getFmgrInfoCmp(cache->info); + cache->elems = cacheAlloc(ctx, sizeof(StatElem) * SPI_processed); + + for(i=0; ivals[i], SPI_tuptable->tupdesc, 1, &isnullvalue); + int64 ndoc; + + if (ndocType == INT4OID) + ndoc = DatumGetInt32(SPI_getbinval(SPI_tuptable->vals[i], SPI_tuptable->tupdesc, 2, &isnullndoc)); + else + ndoc = DatumGetInt64(SPI_getbinval(SPI_tuptable->vals[i], SPI_tuptable->tupdesc, 2, &isnullndoc)); + + if (isnullndoc) + elog(ERROR,"NULL value in second column of table '%s'", tbl); + + if (isnullvalue) + { + /* total number of docs */ + + if (ndoc <= 0) + elog(ERROR,"Total number of document should be positive"); + if ( totaldocs > 0 ) + elog(ERROR,"Total number of document is repeated"); + totaldocs = ndoc; + } + else + { + if ( i>0 && DatumGetInt32( FCall2( &cache->info->cmpFunc, cache->elems[i-1].datum, datum ) ) == 0 ) + elog(ERROR,"Values of first column of table '%s' are not unique", tbl); + + if (ndoc <= 0) + elog(ERROR,"Number of documents with current value should be positive"); + + if ( cache->info->typbyval ) + cache->elems[i].datum = datum; + else + { + size_t size = datumGetSize(datum, false, cache->info->typlen); + + cache->elems[i].datum = PointerGetDatum(cacheAlloc(ctx, size)); + memcpy(DatumGetPointer(cache->elems[i].datum), DatumGetPointer(datum), size); + } + + cache->elems[i].idf = ndoc; + } + } + + if ( totaldocs <= 0) + elog(ERROR,"Total number of document is unknown"); + cache->nelems = SPI_processed - 1; + + for(i=0;inelems;i++) + { + if ( totaldocs < cache->elems[i].idf ) + elog(ERROR,"Inconsitent data in '%s': there is values with frequency > 1", tbl); + cache->elems[i].idf = log( totaldocs / cache->elems[i].idf + getOneAdd() ); + } + } + + SPI_finish(); + + if ( GetSmlarUsePersistent() ) + PersistentDocStat = cache; + + return cache; + } +} + +void +resetStatCache(void) +{ + if ( PersistentDocStat ) + { + + if (!PersistentDocStat->info->typbyval) + { + int i; + for(i=0;inelems;i++) + free( DatumGetPointer(PersistentDocStat->elems[i].datum) ); + } + + if (PersistentDocStat->helems) + free(PersistentDocStat->helems); + free(PersistentDocStat->elems); + free(PersistentDocStat); + } + + PersistentDocStat = NULL; +} + +StatElem * +findStat(StatCache *stat, Datum query, StatElem *low) +{ + StatElem *StopLow = (low) ? low : stat->elems, + *StopHigh = stat->elems + stat->nelems, + *StopMiddle; + int cmp; + + if (stat->info->tupDesc) + elog(ERROR, "TF/IDF is not supported for composite (weighted) type"); + + while (StopLow < StopHigh) + { + StopMiddle = StopLow + ((StopHigh - StopLow) >> 1); + cmp = DatumGetInt32( FCall2( &stat->info->cmpFunc, StopMiddle->datum, query ) ); + + if ( cmp == 0 ) + return StopMiddle; + else if (cmp < 0) + StopLow = StopMiddle + 1; + else + StopHigh = StopMiddle; + } + + return NULL; +} + +void +getHashStatCache(StatCache *stat, MemoryContext ctx, size_t n) +{ + if ( !stat->helems ) + { + stat->helems = cacheAlloc(ctx, (stat->nelems +1) * sizeof(HashedElem)); + stat->selems = cacheAllocZero(ctx, n * sizeof(SignedElem)); + stat->nhelems = -1; + } +} diff --git a/external/smlar/sql/bit.sql b/external/smlar/sql/bit.sql new file mode 100644 index 00000000000..1669561e37b --- /dev/null +++ b/external/smlar/sql/bit.sql @@ -0,0 +1,29 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::int4::bit(10) + FROM + generate_series(1, t) as v + ) AS v + INTO test_bit +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[]) AS s FROM test_bit WHERE v % '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[] ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[]) AS s FROM test_bit WHERE v % '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[] ORDER BY s DESC, t; + +CREATE INDEX idx_test_bit ON test_bit USING gin (v _bit_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[]) AS s FROM test_bit WHERE v % '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[] ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[]) AS s FROM test_bit WHERE v % '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[] ORDER BY s DESC, t; + + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/bytea.sql b/external/smlar/sql/bytea.sql new file mode 100644 index 00000000000..952aa74c5c6 --- /dev/null +++ b/external/smlar/sql/bytea.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::text::bytea + FROM + generate_series(1, t) as v + ) AS v + INTO test_bytea +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_bytea WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_bytea WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_bytea ON test_bytea USING gist (v _bytea_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_bytea WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_bytea WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_bytea; +CREATE INDEX idx_test_bytea ON test_bytea USING gin (v _bytea_sml_ops); + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_bytea WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_bytea WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/char.sql b/external/smlar/sql/char.sql new file mode 100644 index 00000000000..b557da9fefa --- /dev/null +++ b/external/smlar/sql/char.sql @@ -0,0 +1,31 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::varchar(1)::"char" + FROM + generate_series(1, t) as v + ) AS v + INTO test_char +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{3,2,1}') AS s FROM test_char WHERE v % '{3,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_char ON test_char USING gist (v _char_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{3,2,1}') AS s FROM test_char WHERE v % '{3,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_char; +CREATE INDEX idx_test_char ON test_char USING gin (v _char_sml_ops); + +SELECT t, smlar(v, '{3,2,1}') AS s FROM test_char WHERE v % '{3,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/cidr.sql b/external/smlar/sql/cidr.sql new file mode 100644 index 00000000000..d8ff8611ac0 --- /dev/null +++ b/external/smlar/sql/cidr.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + ( '192.168.1.' || (v % 256) )::cidr + FROM + generate_series(1, t) as v + ) AS v + INTO test_cidr +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}') AS s FROM test_cidr WHERE v % '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}') AS s FROM test_cidr WHERE v % '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_cidr ON test_cidr USING gist (v _cidr_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}') AS s FROM test_cidr WHERE v % '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}') AS s FROM test_cidr WHERE v % '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_cidr; +CREATE INDEX idx_test_cidr ON test_cidr USING gin (v _cidr_sml_ops); + +SELECT t, smlar(v, '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}') AS s FROM test_cidr WHERE v % '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}') AS s FROM test_cidr WHERE v % '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/composite_int4.sql b/external/smlar/sql/composite_int4.sql new file mode 100644 index 00000000000..aab069b9623 --- /dev/null +++ b/external/smlar/sql/composite_int4.sql @@ -0,0 +1,46 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +CREATE TYPE cint AS (id int, w float4); +SELECT + t, + ARRAY( + SELECT + (ROW(v::int4, v::float4))::cint + FROM + generate_series(1, t) as v + ) AS v + INTO test_composite_int4 +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[]) AS s + FROM test_composite_int4 + WHERE v % '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[] ORDER BY s DESC, t; +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[] AS s + FROM test_composite_int4 + WHERE v % '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[]) ORDER BY s DESC, t; + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[]) AS s + FROM test_composite_int4 + WHERE smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[]) > 0.6 ORDER BY s DESC, t; +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[]) AS s + FROM test_composite_int4 + WHERE smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[]) > 0.6 ORDER BY s DESC, t; + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[], false) AS s + FROM test_composite_int4 + WHERE smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[], false) > 0.6 ORDER BY s DESC, t; +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[], false) AS s + FROM test_composite_int4 + WHERE smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[], false) > 0.6 ORDER BY s DESC, t; + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[], true) AS s + FROM test_composite_int4 + WHERE smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::cint[], true) > 0.6 ORDER BY s DESC, t; +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[], true) AS s + FROM test_composite_int4 + WHERE smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::cint[], true) > 0.6 ORDER BY s DESC, t; + + diff --git a/external/smlar/sql/composite_text.sql b/external/smlar/sql/composite_text.sql new file mode 100644 index 00000000000..2f41ff81bb8 --- /dev/null +++ b/external/smlar/sql/composite_text.sql @@ -0,0 +1,46 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +CREATE TYPE ctext AS (id text, w float4); +SELECT + t, + ARRAY( + SELECT + (ROW(v::text, v::float4))::ctext + FROM + generate_series(1, t) as v + ) AS v + INTO test_composite_text +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[]) AS s + FROM test_composite_text + WHERE v % '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[] ORDER BY s DESC, t; +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[]) AS s + FROM test_composite_text + WHERE v % '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[] ORDER BY s DESC, t; + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[]) AS s + FROM test_composite_text + WHERE smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[]) > 0.6 ORDER BY s DESC, t; +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[]) AS s + FROM test_composite_text + WHERE smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[]) > 0.6 ORDER BY s DESC, t; + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[], false) AS s + FROM test_composite_text + WHERE smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[], false) > 0.6 ORDER BY s DESC, t; +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[], false) AS s + FROM test_composite_text + WHERE smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[], false) > 0.6 ORDER BY s DESC, t; + +SELECT t, smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[], true) AS s + FROM test_composite_text + WHERE smlar(v, '{"(10,1)", "(9,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(3,1)", "(2,1)", "(1,1)"}'::ctext[], true) > 0.6 ORDER BY s DESC, t; +SELECT t, smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[], true) AS s + FROM test_composite_text + WHERE smlar(v, '{"(50,1)", "(49,1)", "(8,1)", "(7,1)", "(6,1)", "(5,1)", "(4,1)", "(33,1)", "(2,1)", "(1,1)"}'::ctext[], true) > 0.6 ORDER BY s DESC, t; + + diff --git a/external/smlar/sql/date.sql b/external/smlar/sql/date.sql new file mode 100644 index 00000000000..219122c9b8d --- /dev/null +++ b/external/smlar/sql/date.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + epoch2timestamp(v)::date + FROM + generate_series(1, t) as v + ) AS v + INTO test_date +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::date[]) AS s FROM test_date WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::date[] ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::date[]) AS s FROM test_date WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::date[] ORDER BY s DESC, t; + +CREATE INDEX idx_test_date ON test_date USING gist (v _date_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::date[]) AS s FROM test_date WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::date[] ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::date[]) AS s FROM test_date WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::date[] ORDER BY s DESC, t; + +DROP INDEX idx_test_date; +CREATE INDEX idx_test_date ON test_date USING gin (v _date_sml_ops); + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::date[]) AS s FROM test_date WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::date[] ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::date[]) AS s FROM test_date WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::date[] ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/float4.sql b/external/smlar/sql/float4.sql new file mode 100644 index 00000000000..09343c4a5a6 --- /dev/null +++ b/external/smlar/sql/float4.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::float4 + FROM + generate_series(1, t) as v + ) AS v + INTO test_float4 +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_float4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_float4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_float4 ON test_float4 USING gist (v _float4_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_float4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_float4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_float4; +CREATE INDEX idx_test_float4 ON test_float4 USING gin (v _float4_sml_ops); + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_float4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_float4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/float8.sql b/external/smlar/sql/float8.sql new file mode 100644 index 00000000000..3753a42a017 --- /dev/null +++ b/external/smlar/sql/float8.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::float8 + FROM + generate_series(1, t) as v + ) AS v + INTO test_float8 +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_float8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_float8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_float8 ON test_float8 USING gist (v _float8_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_float8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_float8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_float8; +CREATE INDEX idx_test_float8 ON test_float8 USING gin (v _float8_sml_ops); + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_float8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_float8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/inet.sql b/external/smlar/sql/inet.sql new file mode 100644 index 00000000000..7aaa2f138dc --- /dev/null +++ b/external/smlar/sql/inet.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + ( '192.168.1.' || (v % 256) )::inet + FROM + generate_series(1, t) as v + ) AS v + INTO test_inet +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}') AS s FROM test_inet WHERE v % '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}') AS s FROM test_inet WHERE v % '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_inet ON test_inet USING gist (v _inet_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}') AS s FROM test_inet WHERE v % '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}') AS s FROM test_inet WHERE v % '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_inet; +CREATE INDEX idx_test_inet ON test_inet USING gin (v _inet_sml_ops); + +SELECT t, smlar(v, '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}') AS s FROM test_inet WHERE v % '{192.168.1.10,192.168.1.9,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.3,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}') AS s FROM test_inet WHERE v % '{192.168.1.50,192.168.1.49,192.168.1.8,192.168.1.7,192.168.1.6,192.168.1.5,192.168.1.4,192.168.1.33,192.168.1.2,192.168.1.1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/int2.sql b/external/smlar/sql/int2.sql new file mode 100644 index 00000000000..e3e913c7563 --- /dev/null +++ b/external/smlar/sql/int2.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::int2 + FROM + generate_series(1, t) as v + ) AS v + INTO test_int2 +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int2 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int2 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_int2 ON test_int2 USING gist (v _int2_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int2 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int2 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_int2; +CREATE INDEX idx_test_int2 ON test_int2 USING gin (v _int2_sml_ops); + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int2 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int2 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/int4.sql b/external/smlar/sql/int4.sql new file mode 100644 index 00000000000..4a316004b38 --- /dev/null +++ b/external/smlar/sql/int4.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::int4 + FROM + generate_series(1, t) as v + ) AS v + INTO test_int4 +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_int4 ON test_int4 USING gist (v _int4_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_int4; +CREATE INDEX idx_test_int4 ON test_int4 USING gin (v _int4_sml_ops); + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/int4g.sql b/external/smlar/sql/int4g.sql new file mode 100644 index 00000000000..e7d4f09be60 --- /dev/null +++ b/external/smlar/sql/int4g.sql @@ -0,0 +1,143 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.3); + +SET smlar.type = "tfidf"; +SET smlar.stattable = "int4_stat"; + +UPDATE test_int4 SET v = v || '{50,50,50}'::int4[] WHERE t >=50 AND t < 60; +INSERT INTO test_int4 VALUES(201, '{50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50}'); + +CREATE TABLE int4_stat ( + value int4 UNIQUE, + ndoc int4 NOT NULL CHECK (ndoc>0) +); + +INSERT INTO int4_stat ( + SELECT + q.w, + count(*) + FROM + (SELECT array_to_col(v) AS w FROM test_int4 WHERE v IS NOT NULL) AS q + GROUP BY w +); + +INSERT INTO int4_stat VALUES (NULL, (SELECT count(*) FROM test_int4)); + +DROP INDEX idx_test_int4; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_int4 ON test_int4 USING gist (v _int4_sml_ops); +SET enable_seqscan = off; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_int4; +CREATE INDEX idx_test_int4 ON test_int4 USING gin (v _int4_sml_ops); +SET enable_seqscan = off; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int4 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int4 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +RESET enable_seqscan; diff --git a/external/smlar/sql/int4i.sql b/external/smlar/sql/int4i.sql new file mode 100644 index 00000000000..7a9d58d7efd --- /dev/null +++ b/external/smlar/sql/int4i.sql @@ -0,0 +1,15 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(5.0); +SET smlar.type='overlap'; + +DROP INDEX idx_test_int4; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int4[]); +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_int4 WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_int4 ON test_int4 USING gin (v _int4_sml_ops); +SET enable_seqscan = off; + +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_int4 WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + +RESET enable_seqscan; diff --git a/external/smlar/sql/int8.sql b/external/smlar/sql/int8.sql new file mode 100644 index 00000000000..8ea868c01cd --- /dev/null +++ b/external/smlar/sql/int8.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::int8 + FROM + generate_series(1, t) as v + ) AS v + INTO test_int8 +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_int8 ON test_int8 USING gist (v _int8_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_int8; +CREATE INDEX idx_test_int8 ON test_int8 USING gin (v _int8_sml_ops); + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/int8g.sql b/external/smlar/sql/int8g.sql new file mode 100644 index 00000000000..f7262efab12 --- /dev/null +++ b/external/smlar/sql/int8g.sql @@ -0,0 +1,143 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.3); + +SET smlar.type = "tfidf"; +SET smlar.stattable = "int8_stat"; + +UPDATE test_int8 SET v = v || '{50,50,50}'::int8[] WHERE t >=50 AND t < 60; +INSERT INTO test_int8 VALUES(201, '{50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50}'); + +CREATE TABLE int8_stat ( + value int8 UNIQUE, + ndoc int4 NOT NULL CHECK (ndoc>0) +); + +INSERT INTO int8_stat ( + SELECT + q.w, + count(*) + FROM + (SELECT array_to_col(v) AS w FROM test_int8 WHERE v IS NOT NULL) AS q + GROUP BY w +); + +INSERT INTO int8_stat VALUES (NULL, (SELECT count(*) FROM test_int8)); + +DROP INDEX idx_test_int8; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_int8 ON test_int8 USING gist (v _int8_sml_ops); +SET enable_seqscan = off; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_int8; +CREATE INDEX idx_test_int8 ON test_int8 USING gin (v _int8_sml_ops); +SET enable_seqscan = off; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +RESET enable_seqscan; diff --git a/external/smlar/sql/int8i.sql b/external/smlar/sql/int8i.sql new file mode 100644 index 00000000000..b74a464582b --- /dev/null +++ b/external/smlar/sql/int8i.sql @@ -0,0 +1,15 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(5.0); +SET smlar.type='overlap'; + +DROP INDEX idx_test_int8; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_int8 WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_int8 ON test_int8 USING gin (v _int8_sml_ops); +SET enable_seqscan = off; + +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_int8 WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + +RESET enable_seqscan; diff --git a/external/smlar/sql/int8n.sql b/external/smlar/sql/int8n.sql new file mode 100644 index 00000000000..957288ea5a5 --- /dev/null +++ b/external/smlar/sql/int8n.sql @@ -0,0 +1,27 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.8); + +DROP INDEX idx_test_int8; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_int8 ON test_int8 USING gist (v _int8_sml_ops); +SET enable_seqscan = off; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +RESET enable_seqscan; +DROP INDEX idx_test_int8; + +CREATE INDEX idx_test_int8 ON test_int8 USING gin (v _int8_sml_ops); +SET enable_seqscan = off; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::int8[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_int8 WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_int8 WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +RESET enable_seqscan; diff --git a/external/smlar/sql/interval.sql b/external/smlar/sql/interval.sql new file mode 100644 index 00000000000..8611a7c8da2 --- /dev/null +++ b/external/smlar/sql/interval.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::text::interval + FROM + generate_series(1, t) as v + ) AS v + INTO test_interval +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_interval ON test_interval USING gist (v _interval_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_interval; +CREATE INDEX idx_test_interval ON test_interval USING gin (v _interval_sml_ops); + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/intervalg.sql b/external/smlar/sql/intervalg.sql new file mode 100644 index 00000000000..f41394b2bda --- /dev/null +++ b/external/smlar/sql/intervalg.sql @@ -0,0 +1,143 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.3); + +SET smlar.type = "tfidf"; +SET smlar.stattable = "interval_stat"; + +UPDATE test_interval SET v = v || '{50,50,50}'::interval[] WHERE t >=50 AND t < 60; +INSERT INTO test_interval VALUES(201, '{50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50}'); + +CREATE TABLE interval_stat ( + value interval UNIQUE, + ndoc int4 NOT NULL CHECK (ndoc>0) +); + +INSERT INTO interval_stat ( + SELECT + q.w, + count(*) + FROM + (SELECT array_to_col(v) AS w FROM test_interval WHERE v IS NOT NULL) AS q + GROUP BY w +); + +INSERT INTO interval_stat VALUES (NULL, (SELECT count(*) FROM test_interval)); + +DROP INDEX idx_test_interval; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_interval ON test_interval USING gist (v _interval_sml_ops); +SET enable_seqscan = off; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_interval; +CREATE INDEX idx_test_interval ON test_interval USING gin (v _interval_sml_ops); +SET enable_seqscan = off; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +RESET enable_seqscan; diff --git a/external/smlar/sql/intervali.sql b/external/smlar/sql/intervali.sql new file mode 100644 index 00000000000..335ab36a703 --- /dev/null +++ b/external/smlar/sql/intervali.sql @@ -0,0 +1,15 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(5.0); +SET smlar.type='overlap'; + +DROP INDEX idx_test_interval; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_interval WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_interval ON test_interval USING gin (v _interval_sml_ops); +SET enable_seqscan = off; + +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_interval WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + +RESET enable_seqscan; diff --git a/external/smlar/sql/intervaln.sql b/external/smlar/sql/intervaln.sql new file mode 100644 index 00000000000..2e03e0dba91 --- /dev/null +++ b/external/smlar/sql/intervaln.sql @@ -0,0 +1,27 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.8); + +DROP INDEX idx_test_interval; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_interval ON test_interval USING gist (v _interval_sml_ops); +SET enable_seqscan = off; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +RESET enable_seqscan; +DROP INDEX idx_test_interval; + +CREATE INDEX idx_test_interval ON test_interval USING gin (v _interval_sml_ops); +SET enable_seqscan = off; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::interval[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_interval WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_interval WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +RESET enable_seqscan; diff --git a/external/smlar/sql/macaddr.sql b/external/smlar/sql/macaddr.sql new file mode 100644 index 00000000000..9b3635988f2 --- /dev/null +++ b/external/smlar/sql/macaddr.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + ('01:01:01:01:01:' || (v % 100))::macaddr + FROM + generate_series(1, t) as v + ) AS v + INTO test_macaddr +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{01:01:01:01:01:10,01:01:01:01:01:9,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:3,01:01:01:01:01:2,01:01:01:01:01:1}') AS s FROM test_macaddr WHERE v % '{01:01:01:01:01:10,01:01:01:01:01:9,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:3,01:01:01:01:01:2,01:01:01:01:01:1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{01:01:01:01:01:50,01:01:01:01:01:49,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:33,01:01:01:01:01:2,01:01:01:01:01:1}') AS s FROM test_macaddr WHERE v % '{01:01:01:01:01:50,01:01:01:01:01:49,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:33,01:01:01:01:01:2,01:01:01:01:01:1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_macaddr ON test_macaddr USING gist (v _macaddr_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{01:01:01:01:01:10,01:01:01:01:01:9,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:3,01:01:01:01:01:2,01:01:01:01:01:1}') AS s FROM test_macaddr WHERE v % '{01:01:01:01:01:10,01:01:01:01:01:9,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:3,01:01:01:01:01:2,01:01:01:01:01:1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{01:01:01:01:01:50,01:01:01:01:01:49,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:33,01:01:01:01:01:2,01:01:01:01:01:1}') AS s FROM test_macaddr WHERE v % '{01:01:01:01:01:50,01:01:01:01:01:49,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:33,01:01:01:01:01:2,01:01:01:01:01:1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_macaddr; +CREATE INDEX idx_test_macaddr ON test_macaddr USING gin (v _macaddr_sml_ops); + +SELECT t, smlar(v, '{01:01:01:01:01:10,01:01:01:01:01:9,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:3,01:01:01:01:01:2,01:01:01:01:01:1}') AS s FROM test_macaddr WHERE v % '{01:01:01:01:01:10,01:01:01:01:01:9,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:3,01:01:01:01:01:2,01:01:01:01:01:1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{01:01:01:01:01:50,01:01:01:01:01:49,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:33,01:01:01:01:01:2,01:01:01:01:01:1}') AS s FROM test_macaddr WHERE v % '{01:01:01:01:01:50,01:01:01:01:01:49,01:01:01:01:01:8,01:01:01:01:01:7,01:01:01:01:01:6,01:01:01:01:01:5,01:01:01:01:01:4,01:01:01:01:01:33,01:01:01:01:01:2,01:01:01:01:01:1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/money.sql b/external/smlar/sql/money.sql new file mode 100644 index 00000000000..f280e0b1aac --- /dev/null +++ b/external/smlar/sql/money.sql @@ -0,0 +1,28 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::text::money + FROM + generate_series(1, t) as v + ) AS v + INTO test_money +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_money WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_money WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_money ON test_money USING gin (v _money_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_money WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_money WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/numeric.sql b/external/smlar/sql/numeric.sql new file mode 100644 index 00000000000..04626f054ed --- /dev/null +++ b/external/smlar/sql/numeric.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::numeric + FROM + generate_series(1, t) as v + ) AS v + INTO test_numeric +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_numeric WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_numeric WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_numeric ON test_numeric USING gist (v _numeric_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_numeric WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_numeric WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_numeric; +CREATE INDEX idx_test_numeric ON test_numeric USING gin (v _numeric_sml_ops); + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_numeric WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_numeric WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/oid.sql b/external/smlar/sql/oid.sql new file mode 100644 index 00000000000..3c3d5ae34bb --- /dev/null +++ b/external/smlar/sql/oid.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::oid + FROM + generate_series(1, t) as v + ) AS v + INTO test_oid +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_oid WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_oid WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_oid ON test_oid USING gist (v _oid_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_oid WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_oid WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_oid; +CREATE INDEX idx_test_oid ON test_oid USING gin (v _oid_sml_ops); + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_oid WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_oid WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/smlar.sql b/external/smlar/sql/smlar.sql new file mode 100644 index 00000000000..c498675a094 --- /dev/null +++ b/external/smlar/sql/smlar.sql @@ -0,0 +1,136 @@ +CREATE EXTENSION smlar; +set extra_float_digits =0; + +--sanity check +SELECT + opc.opcname, + t.typname, + opc.opcdefault +FROM + pg_opclass opc, + pg_am am, + pg_type t, + pg_type k +WHERE + opc.opcmethod = am.oid AND + am.amname='gist' AND + opc.opcintype = t.oid AND + opc.opckeytype = k.oid AND + k.typname='gsmlsign' +ORDER BY opc.opcname; + +SELECT + opc.opcname, + t.typname, + opc.opcdefault +FROM + pg_opclass opc, + pg_am am, + pg_type t +WHERE + opc.opcmethod = am.oid AND + am.amname='gin' AND + opc.opcintype = t.oid AND + opc.opcname ~ '_sml_ops$' +ORDER BY opc.opcname; + +SELECT + trim( leading '_' from t.typname ) || '[]' AS "Array Type", + gin.opcname AS "GIN operator class", + gist.opcname AS "GiST operator class" +FROM + ( + SELECT * + FROM + pg_catalog.pg_opclass, + pg_catalog.pg_am + WHERE + pg_opclass.opcmethod = pg_am.oid AND + pg_am.amname = 'gin' AND + pg_opclass.opcname ~ '_sml_ops$' + ) AS gin + FULL OUTER JOIN + ( + SELECT * + FROM + pg_catalog.pg_opclass, + pg_catalog.pg_am + WHERE + pg_opclass.opcmethod = pg_am.oid AND + pg_am.amname = 'gist' AND + pg_opclass.opcname ~ '_sml_ops$' + ) AS gist + ON ( + gist.opcname = gin.opcname AND + gist.opcintype = gin.opcintype + ), + pg_catalog.pg_type t +WHERE + t.oid = COALESCE(gist.opcintype, gin.opcintype) AND + t.typarray = 0 +ORDER BY + "Array Type" ASC +; + +SELECT set_smlar_limit(0.1); +SET smlar.threshold = 0.6; + +--First checks +SELECT smlar('{3,2}'::int[], '{3,2,1}'); +SELECT smlar('{2,1}'::int[], '{3,2,1}'); +SELECT smlar('{}'::int[], '{3,2,1}'); +SELECT smlar('{12,10}'::int[], '{3,2,1}'); +SELECT smlar('{123}'::int[], '{}'); +SELECT smlar('{1,4,6}'::int[], '{1,4,6}'); +SELECT smlar('{1,4,6}'::int[], '{6,4,1}'); +SELECT smlar('{1,4,6}'::int[], '{5,4,6}'); +SELECT smlar('{1,4,6}'::int[], '{5,4,6}'); +SELECT smlar('{1,2}'::int[], '{2,2,2,2,2,1}'); +SELECT smlar('{1,2}'::int[], '{1,2,2,2,2,2}'); +SELECT smlar('{}'::int[], '{3}'); +SELECT smlar('{3}'::int[], '{3}'); +SELECT smlar('{2}'::int[], '{3}'); + + +SELECT smlar('{1,4,6}'::int[], '{5,4,6}', 'N.i / (N.a + N.b)' ); +SELECT smlar('{1,4,6}'::int[], '{5,4,6}', 'N.i / sqrt(N.a * N.b)' ); + +SELECT tsvector2textarray('qwe:12 asd:45'); + +SELECT array_unique('{12,12,1,4,1,16}'::int4[]); +SELECT array_unique('{12,12,1,4,1,16}'::bigint[]); +SELECT array_unique('{12,12,1,4,1,16}'::smallint[]); +SELECT array_unique('{12,12,1,4,1,16}'::text[]); +SELECT array_unique('{12,12,1,4,1,16}'::varchar[]); + +SELECT inarray('{12,12,1,4,1,16}'::int[], 13::int); +SELECT inarray('{12,12,1,4,1,16}'::int[], 12::int); +SELECT inarray('{12,12,1,4,1,16}'::text[], 13::text); +SELECT inarray('{12,12,1,4,1,16}'::text[], 12::text); +SELECT inarray('{12,12,1,4,1,16}'::int[], 13::int, 0.9, 0.1); +SELECT inarray('{12,12,1,4,1,16}'::int[], 12::int, 0.9, 0.1); +SELECT inarray('{12,12,1,4,1,16}'::text[], 13::text, 0.9, 0.1); +SELECT inarray('{12,12,1,4,1,16}'::text[], 12::text, 0.9, 0.1); + +--testing function +CREATE OR REPLACE FUNCTION epoch2timestamp(int4) +RETURNS timestamp AS $$ + SELECT ('1970-01-01 00:00:00'::timestamp + ( ($1*3600*24 + $1) ::text || ' seconds' )::interval)::timestamp; + $$ LANGUAGE SQL RETURNS NULL ON NULL INPUT IMMUTABLE; + +CREATE OR REPLACE FUNCTION to_tsp_array(_int4) +RETURNS _timestamp AS $$ + SELECT ARRAY( + SELECT + epoch2timestamp( $1[n] ) + FROM + generate_series( 1, array_upper( $1, 1) - array_lower( $1, 1 ) + 1 ) AS n + ); + $$ LANGUAGE SQL RETURNS NULL ON NULL INPUT IMMUTABLE; + +CREATE OR REPLACE FUNCTION array_to_col(anyarray) +RETURNS SETOF anyelement AS +$$ + SELECT $1[n] FROM generate_series( 1, array_upper( $1, 1) - array_lower( $1, 1 ) + 1 ) AS n; +$$ LANGUAGE SQL RETURNS NULL ON NULL INPUT IMMUTABLE; + diff --git a/external/smlar/sql/text.sql b/external/smlar/sql/text.sql new file mode 100644 index 00000000000..c1609207a57 --- /dev/null +++ b/external/smlar/sql/text.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::text + FROM + generate_series(1, t) as v + ) AS v + INTO test_text +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_text ON test_text USING gist (v _text_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_text; +CREATE INDEX idx_test_text ON test_text USING gin (v _text_sml_ops); + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/textg.sql b/external/smlar/sql/textg.sql new file mode 100644 index 00000000000..e8d2575fb21 --- /dev/null +++ b/external/smlar/sql/textg.sql @@ -0,0 +1,143 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.3); + +SET smlar.type = "tfidf"; +SET smlar.stattable = "text_stat"; + +UPDATE test_text SET v = v || '{50,50,50}'::text[] WHERE t >=50 AND t < 60; +INSERT INTO test_text VALUES(201, '{50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50,50}'); + +CREATE TABLE text_stat ( + value text UNIQUE, + ndoc int4 NOT NULL CHECK (ndoc>0) +); + +INSERT INTO text_stat ( + SELECT + q.w, + count(*) + FROM + (SELECT array_to_col(v) AS w FROM test_text WHERE v IS NOT NULL) AS q + GROUP BY w +); + +INSERT INTO text_stat VALUES (NULL, (SELECT count(*) FROM test_text)); + +DROP INDEX idx_test_text; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_text ON test_text USING gist (v _text_sml_ops); +SET enable_seqscan = off; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_text; +CREATE INDEX idx_test_text ON test_text USING gin (v _text_sml_ops); +SET enable_seqscan = off; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "n"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "log"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = off; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET smlar.tf_method = "const"; +SET smlar.idf_plus_one = on; +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +RESET enable_seqscan; diff --git a/external/smlar/sql/texti.sql b/external/smlar/sql/texti.sql new file mode 100644 index 00000000000..96833e6da80 --- /dev/null +++ b/external/smlar/sql/texti.sql @@ -0,0 +1,15 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(5.0); +SET smlar.type='overlap'; + +DROP INDEX idx_test_text; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_text WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_text ON test_text USING gin (v _text_sml_ops); +SET enable_seqscan = off; + +SELECT t, smlar(v, '{199,198,197,196,195,194,193,192,191,190}') AS s FROM test_text WHERE v % '{199,198,197,196,195,194,193,192,191,190}' ORDER BY s DESC, t; + +RESET enable_seqscan; diff --git a/external/smlar/sql/textn.sql b/external/smlar/sql/textn.sql new file mode 100644 index 00000000000..b3f1ae808b6 --- /dev/null +++ b/external/smlar/sql/textn.sql @@ -0,0 +1,27 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.8); + +DROP INDEX idx_test_text; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_text ON test_text USING gist (v _text_sml_ops); +SET enable_seqscan = off; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +RESET enable_seqscan; +DROP INDEX idx_test_text; + +CREATE INDEX idx_test_text ON test_text USING gin (v _text_sml_ops); +SET enable_seqscan = off; + +SELECT smlar('{199,199,199,199,1}', '{199,1}'::text[]); +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_text WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_text WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +RESET enable_seqscan; diff --git a/external/smlar/sql/time.sql b/external/smlar/sql/time.sql new file mode 100644 index 00000000000..9218a59dbca --- /dev/null +++ b/external/smlar/sql/time.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + epoch2timestamp(v)::time + FROM + generate_series(1, t) as v + ) AS v + INTO test_time +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::time[]) AS s FROM test_time WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::time[] ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::time[]) AS s FROM test_time WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::time[] ORDER BY s DESC, t; + +CREATE INDEX idx_test_time ON test_time USING gist (v _time_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::time[]) AS s FROM test_time WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::time[] ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::time[]) AS s FROM test_time WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::time[] ORDER BY s DESC, t; + +DROP INDEX idx_test_time; +CREATE INDEX idx_test_time ON test_time USING gin (v _time_sml_ops); + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::time[]) AS s FROM test_time WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::time[] ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::time[]) AS s FROM test_time WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::time[] ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/timestamp.sql b/external/smlar/sql/timestamp.sql new file mode 100644 index 00000000000..736e0b84f2d --- /dev/null +++ b/external/smlar/sql/timestamp.sql @@ -0,0 +1,35 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + epoch2timestamp(v)::timestamp + FROM + generate_series(1, t) as v + ) AS v + INTO test_timestamp +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')) AS s FROM test_timestamp WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}') ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')) AS s FROM test_timestamp WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}') ORDER BY s DESC, t; + +CREATE INDEX idx_test_timestamp ON test_timestamp USING gist (v _timestamp_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')) AS s FROM test_timestamp WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}') ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')) AS s FROM test_timestamp WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}') ORDER BY s DESC, t; + +DROP INDEX idx_test_timestamp; +CREATE INDEX idx_test_timestamp ON test_timestamp USING gin (v _timestamp_sml_ops); + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')) AS s FROM test_timestamp WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}') ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')) AS s FROM test_timestamp WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}') ORDER BY s DESC, t; + + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/timestamptz.sql b/external/smlar/sql/timestamptz.sql new file mode 100644 index 00000000000..e2a8687b523 --- /dev/null +++ b/external/smlar/sql/timestamptz.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + epoch2timestamp(v)::timestamptz + FROM + generate_series(1, t) as v + ) AS v + INTO test_timestamptz +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]) AS s FROM test_timestamptz WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[] ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]) AS s FROM test_timestamptz WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[] ORDER BY s DESC, t; + +CREATE INDEX idx_test_timestamptz ON test_timestamptz USING gist (v _timestamptz_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]) AS s FROM test_timestamptz WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[] ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]) AS s FROM test_timestamptz WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[] ORDER BY s DESC, t; + +DROP INDEX idx_test_timestamptz; +CREATE INDEX idx_test_timestamptz ON test_timestamptz USING gin (v _timestamptz_sml_ops); + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]) AS s FROM test_timestamptz WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[] ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]) AS s FROM test_timestamptz WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[] ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/timetz.sql b/external/smlar/sql/timetz.sql new file mode 100644 index 00000000000..8d47bab8e26 --- /dev/null +++ b/external/smlar/sql/timetz.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + epoch2timestamp(v)::timestamptz::timetz + FROM + generate_series(1, t) as v + ) AS v + INTO test_timetz +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]::timetz[]) AS s FROM test_timetz WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]::timetz[] ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]::timetz[]) AS s FROM test_timetz WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]::timetz[] ORDER BY s DESC, t; + +CREATE INDEX idx_test_timetz ON test_timetz USING gist (v _timetz_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]::timetz[]) AS s FROM test_timetz WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]::timetz[] ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]::timetz[]) AS s FROM test_timetz WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]::timetz[] ORDER BY s DESC, t; + +DROP INDEX idx_test_timetz; +CREATE INDEX idx_test_timetz ON test_timetz USING gin (v _timetz_sml_ops); + +SELECT t, smlar(v, to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]::timetz[]) AS s FROM test_timetz WHERE v % to_tsp_array('{10,9,8,7,6,5,4,3,2,1}')::timestamptz[]::timetz[] ORDER BY s DESC, t; +SELECT t, smlar(v, to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]::timetz[]) AS s FROM test_timetz WHERE v % to_tsp_array('{50,49,8,7,6,5,4,33,2,1}')::timestamptz[]::timetz[] ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/varbit.sql b/external/smlar/sql/varbit.sql new file mode 100644 index 00000000000..4389dfea3e0 --- /dev/null +++ b/external/smlar/sql/varbit.sql @@ -0,0 +1,28 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::bit(10)::varbit + FROM + generate_series(1, t) as v + ) AS v + INTO test_varbit +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[]::varbit[]) AS s FROM test_varbit WHERE v % '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[]::varbit[] ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[]::varbit[]) AS s FROM test_varbit WHERE v % '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[]::varbit[] ORDER BY s DESC, t; + +CREATE INDEX idx_test_varbit ON test_varbit USING gin (v _varbit_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[]::varbit[]) AS s FROM test_varbit WHERE v % '{10,9,8,7,6,5,4,3,2,1}'::int4[]::bit(10)[]::varbit[] ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[]::varbit[]) AS s FROM test_varbit WHERE v % '{50,49,8,7,6,5,4,33,2,1}'::int4[]::bit(10)[]::varbit[] ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/sql/varchar.sql b/external/smlar/sql/varchar.sql new file mode 100644 index 00000000000..9faaf32aae8 --- /dev/null +++ b/external/smlar/sql/varchar.sql @@ -0,0 +1,34 @@ +set extra_float_digits =0; +SELECT set_smlar_limit(0.6); + +SELECT + t, + ARRAY( + SELECT + v::varchar + FROM + generate_series(1, t) as v + ) AS v + INTO test_varchar +FROM + generate_series(1, 200) as t; + + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_varchar WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_varchar WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +CREATE INDEX idx_test_varchar ON test_varchar USING gist (v _varchar_sml_ops); + +SET enable_seqscan=off; + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_varchar WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_varchar WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +DROP INDEX idx_test_varchar; +CREATE INDEX idx_test_varchar ON test_varchar USING gin (v _varchar_sml_ops); + +SELECT t, smlar(v, '{10,9,8,7,6,5,4,3,2,1}') AS s FROM test_varchar WHERE v % '{10,9,8,7,6,5,4,3,2,1}' ORDER BY s DESC, t; +SELECT t, smlar(v, '{50,49,8,7,6,5,4,33,2,1}') AS s FROM test_varchar WHERE v % '{50,49,8,7,6,5,4,33,2,1}' ORDER BY s DESC, t; + +SET enable_seqscan=on; + diff --git a/external/smlar/tsarr.c b/external/smlar/tsarr.c new file mode 100644 index 00000000000..4a3deb3360f --- /dev/null +++ b/external/smlar/tsarr.c @@ -0,0 +1,38 @@ +#include "postgres.h" + +#include "catalog/pg_type.h" +#include "tsearch/ts_type.h" +#include "utils/array.h" + +PG_FUNCTION_INFO_V1(tsvector2textarray); +Datum tsvector2textarray(PG_FUNCTION_ARGS); +Datum +tsvector2textarray(PG_FUNCTION_ARGS) +{ + TSVector ts = PG_GETARG_TSVECTOR(0); + ArrayType *a; + Datum *words; + int i; + WordEntry *wptr = ARRPTR(ts); + + words = palloc( sizeof(Datum) * (ts->size+1) ); + + for(i=0; isize; i++) + { + text *t = palloc(VARHDRSZ + wptr->len); + + SET_VARSIZE(t, VARHDRSZ + wptr->len); + memcpy( VARDATA(t), STRPTR(ts) + wptr->pos, wptr->len); + words[i] = PointerGetDatum(t); + + wptr++; + } + + a = construct_array( words, ts->size, + TEXTOID, -1, false, 'i' ); + + PG_FREE_IF_COPY(ts, 0); + + PG_RETURN_ARRAYTYPE_P(a); +} + diff --git a/src/backend/px_optimizer_util/translate/CTranslatorScalarToDXL.cpp b/src/backend/px_optimizer_util/translate/CTranslatorScalarToDXL.cpp index 2c7c3834003..6eca812ea5c 100644 --- a/src/backend/px_optimizer_util/translate/CTranslatorScalarToDXL.cpp +++ b/src/backend/px_optimizer_util/translate/CTranslatorScalarToDXL.cpp @@ -1309,16 +1309,32 @@ CTranslatorScalarToDXL::TranslateArrayCoerceExprToDXL GPOS_ASSERT(NULL != child_node); + Oid elemfuncid = 0; + + if (IsA(array_coerce_expr->elemexpr, FuncExpr)) + { + elemfuncid = ((FuncExpr *) array_coerce_expr->elemexpr)->funcid; + } + else if (IsA(array_coerce_expr->elemexpr, RelabelType)) + { + ; + } + else + { + GPOS_RAISE(gpdxl::ExmaDXL, gpdxl::ExmiQuery2DXLUnsupportedFeature, + GPOS_WSZ_LIT("ArrayCoerceExpr with elemexpr that is neither " + "FuncExpr or RelabelType")); + } + CDXLNode *dxlnode = GPOS_NEW(m_mp) CDXLNode ( m_mp, GPOS_NEW(m_mp) CDXLScalarArrayCoerceExpr ( m_mp, - GPOS_NEW(m_mp) CMDIdGPDB(array_coerce_expr->elemfuncid), + GPOS_NEW(m_mp) CMDIdGPDB(elemfuncid), GPOS_NEW(m_mp) CMDIdGPDB(array_coerce_expr->resulttype), - array_coerce_expr->resulttypmod, - array_coerce_expr->isExplicit, + array_coerce_expr->resulttypmod, true, (EdxlCoercionForm) array_coerce_expr->coerceformat, array_coerce_expr->location ) diff --git a/src/backend/utils/misc/guc.c b/src/backend/utils/misc/guc.c index d5eb35bdd88..8824988bc32 100644 --- a/src/backend/utils/misc/guc.c +++ b/src/backend/utils/misc/guc.c @@ -7875,6 +7875,7 @@ static struct config_string ConfigureNamesString[] = "pg_stat_statements,btree_gin,btree_gist,citext,cube,dict_int,earthdistance," "hstore,intagg,intarray,isn,ltree,pgcrypto,pgrowlocks,pg_prewarm,pg_trgm," "sslinfo,tablefunc,unaccent,fuzzystrmatch,pgstattuple,pg_buffercache,\"uuid-ossp\",bloom," + "smlar," /* POLAR: extensions in src/pl dir */ "plpgsql,plperl," /* PostGIS */ diff --git a/src/include/nodes/primnodes.h b/src/include/nodes/primnodes.h index ccd41ebdd56..47988140b2c 100644 --- a/src/include/nodes/primnodes.h +++ b/src/include/nodes/primnodes.h @@ -884,11 +884,6 @@ typedef struct ArrayCoerceExpr Expr *arg; /* input expression (yields an array) */ Expr *elemexpr; /* expression representing per-element work */ Oid resulttype; /* output type of coercion (an array type) */ - - /* TODO */ - Oid elemfuncid; /* OID of element coercion function, or 0 */ - bool isExplicit; /* conversion semantics flag to pass to func */ - int32 resulttypmod; /* output typmod (also element typmod) */ Oid resultcollid; /* OID of collation, or InvalidOid if none */ CoercionForm coerceformat; /* how to display this node */ diff --git a/src/include/utils/px_sync_guc_name.h b/src/include/utils/px_sync_guc_name.h index 0b64f2461ba..644ab723b0b 100644 --- a/src/include/utils/px_sync_guc_name.h +++ b/src/include/utils/px_sync_guc_name.h @@ -116,6 +116,12 @@ "polar_wait_old_version_page_timeout", "row_security", "search_path", + "smlar.threshold", + "smlar.stattable", + "smlar.type", + "smlar.persistent_cache", + "smlar.idf_plus_one", + "smlar.tf_method", "statement_timeout", "temp_buffers", "temp_tablespaces",