Disable fastpath reading for some data types in ORC (#10939)

Summary:
As discussed in prestodb/presto#23037 (comment), we need to disable fastpath reads of some ORC data types, so that we can add TPCDS related tests in the Presto native module.

CC: Yuhta aditi-pandit

Pull Request resolved: #10939

Reviewed By: Yuhta

Differential Revision: D62373833

Pulled By: mbasmanova

fbshipit-source-id: f38c7959ffb72c1ecbda9c7de4631dfa5ee73e39

velox/dwio/dwrf/reader/SelectiveDecimalColumnReader.h

@@ -33,6 +33,13 @@ class SelectiveDecimalColumnReader : public SelectiveColumnReader {
       DwrfParams& params,
       common::ScanSpec& scanSpec);
 
+  bool hasBulkPath() const override {
+    // Only ORC uses RLEv2 encoding. Currently, ORC decimal data does not
+    // support fastpath reads. When reading RLEv2-encoded decimal data
+    // with null, the query will fail.
+    return version_ != velox::dwrf::RleVersion_2;
+  }
+
   void seekToRowGroup(uint32_t index) override;
   uint64_t skip(uint64_t numValues) override;
 

velox/dwio/dwrf/reader/SelectiveIntegerDirectColumnReader.h

@@ -45,6 +45,7 @@ class SelectiveIntegerDirectColumnReader
     const bool dataVInts = stripe.getUseVInts(si);
 
     format_ = stripe.format();
+    version_ = convertRleVersion(stripe.getEncoding(encodingKey).kind());
     if (format_ == velox::dwrf::DwrfFormat::kDwrf) {
       intDecoder_ = createDirectDecoder</*isSigned=*/true>(
           stripe.getStream(si, params.streamLabels().label(), true),
@@ -64,7 +65,10 @@ class SelectiveIntegerDirectColumnReader
   }
 
   bool hasBulkPath() const override {
-    return true;
+    // Only ORC uses RLEv2 encoding. Currently, ORC integer data does not
+    // support fastpath reads. When reading RLEv2-encoded integer data
+    // with null, the query will fail.
+    return version_ != velox::dwrf::RleVersion_2;
   }
 
   void seekToRowGroup(uint32_t index) override {

velox/dwio/dwrf/reader/SelectiveStringDictionaryColumnReader.h

@@ -32,6 +32,13 @@ class SelectiveStringDictionaryColumnReader
       DwrfParams& params,
       common::ScanSpec& scanSpec);
 
+  bool hasBulkPath() const override {
+    // Only ORC uses RLEv2 encoding. Currently, ORC string data does not
+    // support fastpath reads. When reading RLEv2-encoded string data
+    // with null, the query will fail.
+    return version_ != velox::dwrf::RleVersion_2;
+  }
+
   void seekToRowGroup(uint32_t index) override {
     SelectiveColumnReader::seekToRowGroup(index);
     auto positionsProvider = formatData_->as<DwrfData>().seekToRowGroup(index);

velox/dwio/orc/test/ReaderTest.cpp

@@ -238,3 +238,42 @@ TEST_F(OrcReaderTest, testOrcReadAllType) {
     EXPECT_EQ(structCol->toString(0, 2, ",", false), "{1, 2}");
   }
 }
+
+TEST_F(OrcReaderTest, testOrcRlev2) {
+  google::InstallFailureSignalHandler();
+  const std::string dateOrc(getExamplesFilePath("rlev2.orc"));
+  auto schema =
+      ROW({"id", "price", "name"}, {BIGINT(), DECIMAL(7, 2), VARCHAR()});
+  auto spec = std::make_shared<common::ScanSpec>("<root>");
+  spec->addAllChildFields(*schema);
+
+  dwio::common::ReaderOptions readerOpts{pool()};
+  readerOpts.setScanSpec(spec);
+  readerOpts.setFileFormat(dwio::common::FileFormat::ORC);
+
+  auto reader = DwrfReader::create(
+      createFileBufferedInput(dateOrc, readerOpts.memoryPool()), readerOpts);
+
+  RowReaderOptions rowReaderOptions;
+  rowReaderOptions.setScanSpec(spec);
+  auto rowReader = reader->createRowReader(rowReaderOptions);
+
+  auto batch = BaseVector::create(schema, 0, &readerOpts.memoryPool());
+  while (rowReader->next(500, batch)) {
+    auto rowVector = batch->as<RowVector>();
+    auto idCol =
+        rowVector->childAt(0)->loadedVector()->as<SimpleVector<int64_t>>();
+    auto priceCol =
+        rowVector->childAt(1)->loadedVector()->as<SimpleVector<int64_t>>();
+    auto nameCol =
+        rowVector->childAt(2)->loadedVector()->as<SimpleVector<StringView>>();
+
+    EXPECT_EQ(5, rowVector->size());
+    EXPECT_EQ(idCol->valueAt(0), 1);
+
+    auto priceColType = rowVector->type()->childAt(1);
+    EXPECT_EQ(
+        DecimalUtil::toString(priceCol->valueAt(0), priceColType), "111.11");
+    EXPECT_EQ(nameCol->valueAt(0), "AAAA");
+  }
+}