update

Author: ArnavBalyan

parquet-column/src/main/java/org/apache/parquet/column/page/DataPageV1.java

@@ -88,6 +88,72 @@ public DataPageV1(
     this.indexRowCount = rowCount;
   }
 
+  /**
+   * Constructor with explicit compressed size.
+   * Use this when you know both compressed and uncompressed sizes.
+   *
+   * @param bytes            the bytes for this page (compressed or decompressed)
+   * @param valueCount       count of values in this page
+   * @param compressedSize   the actual compressed size of the page
+   * @param uncompressedSize the uncompressed size of the page
+   * @param statistics       of the page's values (max, min, num_null)
+   * @param rlEncoding       the repetition level encoding for this page
+   * @param dlEncoding       the definition level encoding for this page
+   * @param valuesEncoding   the values encoding for this page
+   */
+  public DataPageV1(
+      BytesInput bytes,
+      int valueCount,
+      int compressedSize,
+      int uncompressedSize,
+      Statistics<?> statistics,
+      Encoding rlEncoding,
+      Encoding dlEncoding,
+      Encoding valuesEncoding) {
+    super(compressedSize, uncompressedSize, valueCount);
+    this.bytes = bytes;
+    this.statistics = statistics;
+    this.rlEncoding = rlEncoding;
+    this.dlEncoding = dlEncoding;
+    this.valuesEncoding = valuesEncoding;
+    this.indexRowCount = -1;
+  }
+
+  /**
+   * Constructor with explicit compressed size and row info.
+   * Use this when you know both compressed and uncompressed sizes plus row information.
+   *
+   * @param bytes
+   * @param valueCount
+   * @param compressedSize
+   * @param uncompressedSize
+   * @param firstRowIndex
+   * @param rowCount
+   * @param statistics
+   * @param rlEncoding
+   * @param dlEncoding
+   * @param valuesEncoding
+   */
+  public DataPageV1(
+      BytesInput bytes,
+      int valueCount,
+      int compressedSize,
+      int uncompressedSize,
+      long firstRowIndex,
+      int rowCount,
+      Statistics<?> statistics,
+      Encoding rlEncoding,
+      Encoding dlEncoding,
+      Encoding valuesEncoding) {
+    super(compressedSize, uncompressedSize, valueCount, firstRowIndex);
+    this.bytes = bytes;
+    this.statistics = statistics;
+    this.rlEncoding = rlEncoding;
+    this.dlEncoding = dlEncoding;
+    this.valuesEncoding = valuesEncoding;
+    this.indexRowCount = rowCount;
+  }
+
   /**
    * @return the bytes for the page
    */

parquet-hadoop/src/main/java/org/apache/parquet/hadoop/ColumnChunkPageReadStore.java

@@ -184,6 +184,7 @@ public DataPage visit(DataPageV1 dataPageV1) {
               decompressedPage = new DataPageV1(
                   decompressed,
                   dataPageV1.getValueCount(),
+                  dataPageV1.getCompressedSize(),
                   dataPageV1.getUncompressedSize(),
                   dataPageV1.getStatistics(),
                   dataPageV1.getRlEncoding(),
@@ -194,6 +195,7 @@ public DataPage visit(DataPageV1 dataPageV1) {
               decompressedPage = new DataPageV1(
                   decompressed,
                   dataPageV1.getValueCount(),
+                  dataPageV1.getCompressedSize(),
                   dataPageV1.getUncompressedSize(),
                   firstRowIndex,
                   Math.toIntExact(offsetIndex.getLastRowIndex(currentPageIndex, rowCount)

parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestDataPageChecksums.java

@@ -50,6 +50,7 @@
 import org.apache.parquet.column.page.DictionaryPage;
 import org.apache.parquet.column.page.Page;
 import org.apache.parquet.column.page.PageReadStore;
+import org.apache.parquet.column.page.PageReader;
 import org.apache.parquet.column.page.PageWriter;
 import org.apache.parquet.column.statistics.Statistics;
 import org.apache.parquet.compression.CompressionCodecFactory.BytesInputCompressor;
@@ -771,4 +772,35 @@ private void assertVerificationFailed(ParquetFileReader reader) {
           e.getMessage().contains("CRC checksum verification failed"));
     }
   }
+
+  @Test
+  public void testCompressedSizePreservedAfterDecompression() throws IOException {
+    Configuration conf = new Configuration();
+
+    CompressionCodecName compression = CompressionCodecName.GZIP;
+    ParquetProperties.WriterVersion version = ParquetProperties.WriterVersion.PARQUET_1_0;
+
+    Path testFile = writeSimpleParquetFile(conf, compression, version);
+
+    try (ParquetFileReader reader = ParquetFileReader.open(HadoopInputFile.fromPath(testFile, conf))) {
+      PageReadStore pageStore = reader.readNextRowGroup();
+
+      ColumnDescriptor colDesc =
+          reader.getFileMetaData().getSchema().getColumns().get(0);
+      PageReader pageReader = pageStore.getPageReader(colDesc);
+
+      DataPage page;
+      while ((page = pageReader.readPage()) != null) {
+        int compressedSize = page.getCompressedSize();
+        int uncompressedSize = page.getUncompressedSize();
+
+        assertTrue(
+            "Compressed size (" + compressedSize + ") should be less than uncompressed size ("
+                + uncompressedSize + ") for compressed pages",
+            compressedSize < uncompressedSize);
+
+        assertTrue("Compressed size should be positive", compressedSize > 0);
+      }
+    }
+  }
 }