From ede6a69dd0a54c088fd4d91722793c3ed9d775bb Mon Sep 17 00:00:00 2001
From: kevincmchen <kc4163568@gmail.com>
Date: Wed, 31 Jul 2024 19:21:38 -0700
Subject: [PATCH] support to read fixed length binary as string. (#10621)

Summary:
This is a followup of https://github.com/facebookincubator/velox/issues/10399,
support to read fixed length binary as string.

Pull Request resolved: https://github.com/facebookincubator/velox/pull/10621

Reviewed By: Yuhta

Differential Revision: D60527079

Pulled By: kagamiori

fbshipit-source-id: a13f3bb963c57e98d494fb9fc194f2a988ba2bee
---
 velox/dwio/parquet/reader/PageReader.cpp      |   2 +-
 .../dwio/parquet/tests/examples/uuid.parquet  | Bin 0 -> 730 bytes
 .../tests/reader/ParquetReaderTest.cpp        |  28 ++++++++++++++++++
 3 files changed, 29 insertions(+), 1 deletion(-)
 create mode 100644 velox/dwio/parquet/tests/examples/uuid.parquet

diff --git a/velox/dwio/parquet/reader/PageReader.cpp b/velox/dwio/parquet/reader/PageReader.cpp
index 004b0f6b801c..cf46fdb58184 100644
--- a/velox/dwio/parquet/reader/PageReader.cpp
+++ b/velox/dwio/parquet/reader/PageReader.cpp
@@ -666,7 +666,7 @@ void PageReader::makeDecoder() {
               pageData_, pageData_ + encodedDataSize_);
           break;
         case thrift::Type::FIXED_LEN_BYTE_ARRAY:
-          if (type_->type()->isVarbinary()) {
+          if (type_->type()->isVarbinary() || type_->type()->isVarchar()) {
             stringDecoder_ = std::make_unique<StringDecoder>(
                 pageData_, pageData_ + encodedDataSize_, type_->typeLength_);
           } else {
diff --git a/velox/dwio/parquet/tests/examples/uuid.parquet b/velox/dwio/parquet/tests/examples/uuid.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..91ca9d2061fd979c7822c1833e80cdec68fd3424
GIT binary patch
literal 730
zcma))&5G1O5XU=Fh9J1GXoo~{7{WTNFddSvB%Lf?)`KkKVcDyQbo#>$&d1u5Ssj;o
z0r9X$-@u!XA)Y;X@*qBdk6>s0SUm`H>#F)w_4`-#;r>yE5&rDpFAo0n@$0+m5gyot
zhdqQQ-=7tES>;8}36!9T5TYVXrVU|b$|z@r)SOewin_=(fm+w3fifdH6%CP;<|5Cy
zW>BQSKyVJEC{;zIC@P|rVnl*SE0$8mD+ssUFxc5h8SHEVhTYIdD2#j$pS$*jY+`(H
z>wf!E$%kt8v^&1H>eRk5e08M6+(_&qyoC=i3a%}et$d<eHIc|4gi#ocya;cFw>=d8
zUBs?G@L#+BxjXXimS}e)uFXtZaC->LjFA7~3ikcCE`INze%fH<*@-u;w{dXegq!C}
z)g}DY%#(#UR#P6lan9myb)w?E*jzY^$MK9$`}qIvt>(h;BHnv+_FI0r2V4Hx@=t5s
zFY>bUKhj0}%1(oFy?(q#!S)N&cFH8vxm1&+Ckd|hyeN3^3!hBQ@Kl*aJD&}+B%=u(
w?l#MIB8QAANJW*_09C0(K?MlSO96Sqlm?&~P%7&DP8db@#@`%-4(+f00m_lWlmGw#

literal 0
HcmV?d00001

diff --git a/velox/dwio/parquet/tests/reader/ParquetReaderTest.cpp b/velox/dwio/parquet/tests/reader/ParquetReaderTest.cpp
index 1a4ef241bd3c..743f14682627 100644
--- a/velox/dwio/parquet/tests/reader/ParquetReaderTest.cpp
+++ b/velox/dwio/parquet/tests/reader/ParquetReaderTest.cpp
@@ -1180,6 +1180,34 @@ TEST_F(ParquetReaderTest, readBinaryAsStringFromNation) {
           0));
 }
 
+TEST_F(ParquetReaderTest, readFixedLenBinaryAsStringFromUuid) {
+  const std::string filename("uuid.parquet");
+  const std::string sample(getExampleFilePath(filename));
+
+  dwio::common::ReaderOptions readerOptions{leafPool_.get()};
+  auto outputRowType = ROW({"uuid_field"}, {VARCHAR()});
+
+  readerOptions.setFileSchema(outputRowType);
+  auto reader = createReader(sample, readerOptions);
+  EXPECT_EQ(reader->numberOfRows(), 3ULL);
+  auto rowType = reader->typeWithId();
+  EXPECT_EQ(rowType->type()->kind(), TypeKind::ROW);
+  EXPECT_EQ(rowType->size(), 1ULL);
+  EXPECT_EQ(rowType->childAt(0)->type()->kind(), TypeKind::VARCHAR);
+
+  auto rowReaderOpts = getReaderOpts(outputRowType);
+  rowReaderOpts.setScanSpec(makeScanSpec(outputRowType));
+  auto rowReader = reader->createRowReader(rowReaderOpts);
+
+  auto expected = std::string("5468454a-363f-ccc8-7d0b-76072a75dfaa");
+  VectorPtr result = BaseVector::create(outputRowType, 0, &(*leafPool_));
+  rowReader->next(1, result);
+  EXPECT_EQ(
+      expected,
+      result->as<RowVector>()->childAt(0)->asFlatVector<StringView>()->valueAt(
+          0));
+}
+
 TEST_F(ParquetReaderTest, testV2PageWithZeroMaxDefRep) {
   // enum_type.parquet contains 1 column (ENUM) with 3 rows.
   const std::string sample(getExampleFilePath("v2_page.parquet"));