Set parquet version better in writer

[gaborcsardi]

Versions needed for some features:

2.0

Fri Dec 13 16:44:07 2013
574ac1903190c253cd7a6716c6c9bb2b9ae0f4a1
tag: parquet-format-2.0.0

• sorted row groups: https://apache.googlesource.com/parquet-format/+/52af89e5aefc6c9290b38350cd060517725645ff%5E1..52af89e5aefc6c9290b38350cd060517725645ff/
• DELTA_LENGTH_BYTE_ARRAY: https://apache.googlesource.com/parquet-format/+/36d1a0cc622e9152d21830bab2e668990c35a3b4%5E1..36d1a0cc622e9152d21830bab2e668990c35a3b4/
• Statistics: https://apache.googlesource.com/parquet-format/+/2c4ada8e988c1c0018332e420764c09bab16bf9c
• data page v2: https://apache.googlesource.com/parquet-format/+/227cef3c992c8edbf929134f86aaeed1e2f54bd5
• ENUM converted type: https://apache.googlesource.com/parquet-format/+/3aa2c134bd5210a59763f5e64517ad2b75d2639d
• DELTA_BINARY_PACKED encoding: https://apache.googlesource.com/parquet-format/+/9a38528e47760912f73ecebae262596d51c12463 (+ others)
• RLE_DICTIONARY encoding! https://apache.googlesource.com/parquet-format/+/5ab14d91c5364a2eca80882a6ea2bf65bc051051%5E%21/ (before this, really)
• is_sorted flag for dictionary pages. (This is optional, does that mean that older thrift will still read it?): https://apache.googlesource.com/parquet-format/+/3bc09d96c832a633777a0099f9473fa4c91e2c8f
• DELTA_BYTE_ARRAY encoding: https://apache.googlesource.com/parquet-format/+/94d23109041ee6b55a5d8a049663e4128fed63c4 (this is just the name, earlier, really)

2.1

• field_id in SchemaElement: https://apache.googlesource.com/parquet-format/+/d456669c63697e61db05e883526324880a502d90
• statistics in v1 data pages: https://apache.googlesource.com/parquet-format/+/89c85601b1ef19402b9474c4f31417088dabc917
• DECIMAL converted type, scale and precision in SchemaElement: https://apache.googlesource.com/parquet-format/+/b2836e591da8216cfca47075baee2c9a7b0b9289

2.2

• DATE converted type: https://apache.googlesource.com/parquet-format/+/090444ca7315651791f108755955663dfab7c942
• TIME_MILLIS converted type: https://apache.googlesource.com/parquet-format/+/090444ca7315651791f108755955663dfab7c942
• TIMESTAMP_MILLIS, UNIT_*, INT_*, JSON, BSON, INTERVAL converted types: https://apache.googlesource.com/parquet-format/+/090444ca7315651791f108755955663dfab7c942
• PageEncodingStats in ColumnMetaData: https://apache.googlesource.com/parquet-format/+/f7ab552f569df63bdb59f751d0dd36e826682739

2.3

Nothing changed in the format?

2.4

Tue Oct 17 12:25:18 2017
08eb0cee5b5fdb1bc88cd684448be123754b8c06
apache-parquet-format-2.4.0

• TIME_MICROS, TIMESTAMP_MICROS converted types: https://apache.googlesource.com/parquet-format/+/5b806d1e855bf47f5234c768aefc000b704f43ab
• BROTLI compression: https://apache.googlesource.com/parquet-format/+/66a5a7b982e291e06afb1da7ffe9da211318caba
• min_value and max_value in Statistics: https://apache.googlesource.com/parquet-format/+/041708da1af52e7cb9288c331b542aa25b68a2b6
• LogicalType: https://apache.googlesource.com/parquet-format/+/863875e0be3237c6aa4ed71733d54c91a51deabe
• UUID logical type: https://apache.googlesource.com/parquet-format/+/ddc18a7af21127f9100096b5b356d1cad888d174
• lz4 and zstd compression: https://apache.googlesource.com/parquet-format/+/84460c5a1e8aadf52a40dcf2aeb2fc875df4ac2a
• Column indices: https://apache.googlesource.com/parquet-format/+/f1de77d31936f4d50f1286676a0034b6339918ee

2.5

Some deprecations, but no changes.

2.6

• The format didn't include UUID logical type before: https://apache.googlesource.com/parquet-format/+/3e6cd1415eabdd8ec93881a5c89ae82fb150ab29
• file_offset and total_compressed_size in RowGroup: https://apache.googlesource.com/parquet-format/+/677ed8ea23c60e5e42a4c537a454544884525593
• nanoseconds precision time units: https://apache.googlesource.com/parquet-format/+/b879065ac1bee3fe1d770eb3c4b60ab4267044d7

2.7

Sep 25 09:18:41 2019
ee5cae066ed602bd969024eb308c5262c451b6cd
apache-parquet-format-2.7.0

• bloom filters: https://apache.googlesource.com/parquet-format/+/54839ad5e04314c944fed8aa4bc6cf15e4a58698 + a lot more
• encryption: https://apache.googlesource.com/parquet-format/+/518e206c3e6586b76e8315d5f62a8666ed62fa90

2.8

Dec 16 14:21:26 2019
66cb393a0f4fe93cddc49345cc2a45490ea234e1
apache-parquet-format-2.8.0

• BYTE_STREAM_SPLIT encoding: https://apache.googlesource.com/parquet-format/+/ee02ef8c8f33bd3d5ed0582ded7e20439e12d933

2.9

Apr 7 14:50:21 2021
b4f0c0a643a6ec1a7def37115dd6967ba9346df7
apache-parquet-format-2.9.0

• LZ4_RAW compression: https://apache.googlesource.com/parquet-format/+/7f06e838cbd1b7dbd722ff2580b9c2525e37fc46

2.10

Nov 16 12:36:59 2023
b9c4fa81c3be13dc98760c92b037fa4dd465cef8
apache-parquet-format-2.10

• DELTA_BYTE_ARRAY for FIXED_LEN_BYTE_ARRAY (but some tools were already supporting it): https://apache.googlesource.com/parquet-format/+/bfc549b93e6927cb1fc425466e4084f76edc6d22
• bloom_filter_length in ColumnMetaData: https://apache.googlesource.com/parquet-format/+/9b040ccbec388a5e664976ea8dda7b60bde02685
• is_max_value_exact and is_min_value_exact in Statistics: https://apache.googlesource.com/parquet-format/+/31f92c73ecca63b596b5edb391f9ac5eba9dbbf8
• FLOAT16 logical type: https://apache.googlesource.com/parquet-format/+/46cc3a0647d301bb9579ca8dd2cc356caf2a72d2
• size_statistics in ColumnMetaData: https://apache.googlesource.com/parquet-format/+/40699d05bd24181de6b1457babbee2c16dce3803

2.11 (unreleased)

• Extend BYTE_STREAM_SPLIT to support INT32, INT64 and FIXED_LEN_BYTE_ARRAY data: https://apache.googlesource.com/parquet-format/+/e517ac4dbe08d518eb5c2e58576d4c711973db94
• Deprecate ColumnChunk::file_offset field: https://apache.googlesource.com/parquet-format/+/5b564f3c47679526cf72e54f207013f28f53acc4
• VariantType