vrozov commented on a change in pull request #1428: DRILL-6670: align Parquet
TIMESTAMP_MICROS logical type handling with earlier versions + minor fixes
URL: https://github.com/apache/drill/pull/1428#discussion_r209014522
##########
File path:
exec/java-exec/src/main/java/org/apache/drill/exec/store/parquet/AbstractParquetScanBatchCreator.java
##########
@@ -108,7 +108,7 @@ protected ScanBatch getBatch(ExecutorFragmentContext
context, AbstractParquetRow
if
(!context.getOptions().getBoolean(ExecConstants.PARQUET_NEW_RECORD_READER)
&& !ParquetReaderUtility.containsComplexColumn(footer,
rowGroupScan.getColumns())) {
- logger.debug("Query {} qualifies for new Parquet reader",
+ logger.debug("Query {} qualifies for ParquetRecordReader",
Review comment:
IMO, it will be much better to log `PARQUET_NEW_RECORD_READER` option and
whether or not there are complex columns prior to creating a reader and reader
instance class after the reader is created:
```
log.debug("PARQUET_NEW_RECORD_READER is {}. Complex columns {}.",
"enabled"/"disabled", "found"/"not found");
Reader reader;
if (...) {
reader = new ParquetRecordReader()l
} else {
reader = new DrillParquetReader();
}
log.debug("Query {} uses {});
readers.add(reader);
```
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
