Github user saucam commented on a diff in the pull request:
https://github.com/apache/spark/pull/2841#discussion_r19410910
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/parquet/ParquetTableOperations.scala
---
@@ -450,8 +465,14 @@ private[parquet] class FilteringParquetRowInputFormat
globalMetaData.getKeyValueMetaData(),
globalMetaData.getSchema()))
+ val filter: Filter = ParquetInputFormat.getFilter(configuration)
+ var rowGroupsDropped :Long = 0
+ var totalRowGroups :Long = 0
+
+ // Ugly hack, stuck with it until mentioned PR is resolved
val generateSplits =
- classOf[ParquetInputFormat[_]].getDeclaredMethods.find(_.getName ==
"generateSplits").get
+ Class.forName("parquet.hadoop.ClientSideMetadataSplitStrategy")
--- End diff --
Thanks for pointing this out, I had chosen ClientSideMetadataStrategy
because it was the default one (had left the other as TODO), in fact we indeed
should use TaskSideMetadataSplitStrategy . For now I have included the code
for both strategies, and made TaskSideMetadataSplitStrategy default. Should we
create a config for this ?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
