maryannxue commented on a change in pull request #27742: [SPARK-30991] Refactor
AQE readers and RDDs
URL: https://github.com/apache/spark/pull/27742#discussion_r385997172
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/adaptive/ReduceNumShufflePartitions.scala
##########
@@ -36,16 +36,15 @@ case class ReduceNumShufflePartitions(conf: SQLConf)
extends Rule[SparkPlan] {
if (!conf.reducePostShufflePartitionsEnabled) {
return plan
}
- if (!plan.collectLeaves().forall(_.isInstanceOf[QueryStageExec])) {
+ if (!plan.collectLeaves().forall(_.isInstanceOf[QueryStageExec])
+ || plan.find(_.isInstanceOf[CustomShuffleReaderExec]).isDefined) {
Review comment:
@cloud-fan Is that we should skip this optimization at all if there's a skew
join instead of skip counting the stages as in
https://github.com/apache/spark/pull/27742/files#diff-e23b4656e59b73d313271d62329eefc2L47-L48
?
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
