fix: Throw error for unsupported compression kinds in TextReader (facebookincubator#15316)

Summary:
This PR fixes an issue where TextReader would produce garbled output when
trying to read files with unsupported compression formats (like SNAPPY,
LZO, LZ4). Instead of failing silently, TextReader will now check the
compression kind upon initialization and throw an explicit exception if
it is not supported. This prevents data corruption and provides clear
feedback to the user.

This PR fixes facebookincubator#15314.

Pull Request resolved: facebookincubator#15316

Reviewed By: kevinwilfong

Differential Revision: D85881726

Pulled By: zacw7

fbshipit-source-id: b08920b030dcc5d50ac1b9bcfb0a67a6fdd1b417

Author: xiaodouchen

velox/dwio/text/reader/TextReader.cpp

@@ -35,6 +35,9 @@ using folly::StringPiece;
 constexpr const char* kTextfileCompressionExtensionGzip = ".gz";
 constexpr const char* kTextfileCompressionExtensionDeflate = ".deflate";
 constexpr const char* kTextfileCompressionExtensionZst = ".zst";
+constexpr const char* kTextfileCompressionExtensionLz4 = ".lz4";
+constexpr const char* kTextfileCompressionExtensionLzo = ".lzo";
+constexpr const char* kTextfileCompressionExtensionSnappy = ".snappy";
 
 static std::string emptyString = std::string();
 
@@ -101,6 +104,11 @@ void setCompressionSettings(
     const std::string& filename,
     CompressionKind& kind,
     dwio::common::compression::CompressionOptions& compressionOptions) {
+  if (endsWith(filename, kTextfileCompressionExtensionLz4) ||
+      endsWith(filename, kTextfileCompressionExtensionLzo) ||
+      endsWith(filename, kTextfileCompressionExtensionSnappy)) {
+    VELOX_FAIL("Unsupported compression extension for file: {}", filename);
+  }
   if (endsWith(filename, kTextfileCompressionExtensionGzip)) {
     kind = CompressionKind::CompressionKind_GZIP;
     compressionOptions.format.zlib.windowBits =

velox/dwio/text/tests/reader/TextReaderTest.cpp

@@ -1910,6 +1910,31 @@ INSTANTIATE_TEST_SUITE_P(
     testing::ValuesIn(params),
     [](const auto& paramInfo) { return paramInfo.param.compression; });
 
+TEST_F(TextReaderTest, unsupportedCompressedKind) {
+  auto type = ROW(
+      {{"col_string", VARCHAR()},
+       {"col_int", INTEGER()},
+       {"col_float", DOUBLE()},
+       {"col_bool", BOOLEAN()}});
+  auto factory = dwio::common::getReaderFactory(dwio::common::FileFormat::TEXT);
+  const std::string kBaseDir = "velox/dwio/text/tests/reader/";
+  std::vector paths = {
+      getDataFilePath(kBaseDir, "examples/simple_types_compressed_file.lz4"),
+      getDataFilePath(kBaseDir, "examples/simple_types_compressed_file.lzo"),
+      getDataFilePath(
+          kBaseDir, "examples/simple_types_compressed_file.snappy")};
+  for (const auto& path : paths) {
+    auto readFile = std::make_shared<LocalReadFile>(path);
+    auto readerOptions = dwio::common::ReaderOptions(pool());
+    readerOptions.setFileSchema(type);
+    auto input =
+        std::make_unique<dwio::common::BufferedInput>(readFile, poolRef());
+    EXPECT_THROW(
+        factory->createReader(std::move(input), readerOptions),
+        VeloxRuntimeError);
+  }
+}
+
 } // namespace
 
 } // namespace facebook::velox::text