pitrou commented on code in PR #3368:
URL: https://github.com/apache/parquet-java/pull/3368#discussion_r2597962565
##########
parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestColumnChunkPageWriteStore.java:
##########
@@ -316,7 +316,130 @@ public void testColumnOrderV1() throws IOException {
}
}
+ @Test
+ public void testV2PageCompressThreshold() throws Exception {
+ allocator = TrackingByteBufferAllocator.wrap(new
HeapByteBufferAllocator());
+ MessageType schema = MessageTypeParser.parseMessageType("message test {
required int32 data; }");
+ ColumnDescriptor col = schema.getColumns().get(0);
+
+ int valueCount = 100;
+ int rowCount = 100;
+ int nullCount = 0;
+ ByteArrayOutputStream baos = new ByteArrayOutputStream();
+ byte[] value =
+
ByteBuffer.allocate(4).order(ByteOrder.LITTLE_ENDIAN).putInt(0).array();
+ for (int i = 0; i < valueCount; i++) {
+ baos.write(value);
+ }
+ BytesInput data = BytesInput.from(baos.toByteArray());
+
+ BytesInput definitionLevels = BytesInput.empty();
+ BytesInput repetitionLevels = BytesInput.empty();
+ Statistics<?> statistics = Statistics.getBuilderForReading(
+ Types.required(INT32).named("data"))
+ .build();
+ statistics.setMinMaxFromBytes(value, value);
+
+ Path file = writeSingleColumnData(
+ "lowCompressThreshold",
+ schema,
+ col,
+ rowCount,
+ nullCount,
+ valueCount,
+ repetitionLevels,
+ definitionLevels,
+ data,
+ statistics,
+ 0.01);
+ ParquetMetadata footer = ParquetFileReader.readFooter(conf, file,
NO_FILTER);
+ ColumnChunkMetaData columnMeta =
footer.getBlocks().get(0).getColumns().get(0);
+
+ long uncompressedSize = columnMeta.getTotalUncompressedSize();
+ long compressedSize = columnMeta.getTotalSize();
+
+ assertEquals(
+ "Data should be stored uncompressed when compression ratio exceeds
threshold",
+ uncompressedSize,
+ compressedSize);
+
+ file = writeSingleColumnData(
+ "highCompressThreshold",
+ schema,
+ col,
+ rowCount,
+ nullCount,
+ valueCount,
+ repetitionLevels,
+ definitionLevels,
+ data,
+ statistics,
+ 10);
Review Comment:
It does not make sense to have a threshold greater than 1.0, does it?
In any case, the data of a column full of zeros is extremely compressible
(you might want to disable dictionary encoding?), so the default value should
work IMHO.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
