faucct commented on code in PR #37749:
URL: https://github.com/apache/spark/pull/37749#discussion_r1430108415
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/physical/partitioning.scala:
##########
@@ -361,6 +361,18 @@ object KeyGroupedPartitioning {
partitionValues: Seq[InternalRow]): KeyGroupedPartitioning = {
KeyGroupedPartitioning(expressions, partitionValues.size,
Some(partitionValues))
}
+
+ def supportsExpressions(expressions: Seq[Expression]): Boolean = {
Review Comment:
I guess, this code is supposed to make sure there is only one child, but it
was skipped in my execution as `sorted` was empty, though I don't understand
what is this requirement for:
https://github.com/apache/spark/pull/35657/files#diff-715d0c2d59a4ddb8a4b5952c1b05be9f035b6d9b0d9670c70b58989dc722b252R86
Making `sorted` empty let me pass this check, though I need it for the join
optimization. But I have managed to bring this property back on the physical
level using `org.apache.spark.sql.connector.read.SupportsReportOrdering` and
execute the `SortMergeJoin` without the exchanges and sorts.
--
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]
