huaxiangsun commented on code in PR #12771:
URL: https://github.com/apache/iceberg/pull/12771#discussion_r2039688419
##########
spark/v3.5/spark/src/test/java/org/apache/iceberg/spark/data/TestSparkParquetWriter.java:
##########
@@ -151,4 +152,27 @@ public void testFpp() throws IOException,
NoSuchFieldException, IllegalAccessExc
assertThat(fpp).isEqualTo(0.05);
}
}
+
+ @Test
+ public void testColumnStatsEnabled()
+ throws IOException, NoSuchFieldException, IllegalAccessException {
+ File testFile = File.createTempFile("junit", null, temp.toFile());
+ try (FileAppender<InternalRow> writer =
+ Parquet.write(Files.localOutput(testFile))
+ .schema(SCHEMA)
+ .set(PARQUET_COLUMN_STATS_ENABLED_PREFIX + "id_long", "false")
+ .createWriterFunc(
+ msgType ->
+
SparkParquetWriters.buildWriter(SparkSchemaUtil.convert(SCHEMA), msgType))
+ .build()) {
+ // Using reflection to access the private 'props' field in ParquetWriter
+ Field propsField = writer.getClass().getDeclaredField("props");
+ propsField.setAccessible(true);
+ ParquetProperties props = (ParquetProperties) propsField.get(writer);
+ MessageType parquetSchema = ParquetSchemaUtil.convert(SCHEMA, "test");
+ ColumnDescriptor idlDescriptor = parquetSchema.getColumnDescription(new
String[] {"id_long"});
+ // Default statisticsEnabled should be true and for column id_long, it
is disabled.
+ assertThat(props.getStatisticsEnabled(idlDescriptor)).isEqualTo(false);
+ }
+ }
Review Comment:
Thanks @pvary for review. I will read more to see if the test can be moved
to TestParquet and if test on the parquet files.
As for default, there is a bug in Parquet-mr module with default,
https://github.com/apache/parquet-java/pull/3189, right now, the default is
always true which means it cannot be really configured.
My intention is that:
1. User can configure default to true, and disable stats for specific
columns. or
2. User can configure default to false, enable stats for specific columns.
--
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]
