jackylee-ch commented on code in PR #44661:
URL: https://github.com/apache/spark/pull/44661#discussion_r1455516769
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/adaptive/CoalesceShufflePartitions.scala:
##########
@@ -146,13 +147,15 @@ case class CoalesceShufflePartitions(session:
SparkSession) extends AQEShuffleRe
Seq(collectShuffleStageInfos(r))
case unary: UnaryExecNode => collectCoalesceGroups(unary.child)
case union: UnionExec => union.children.flatMap(collectCoalesceGroups)
- // If not all leaf nodes are exchange query stages, it's not safe to
reduce the number of
- // shuffle partitions, because we may break the assumption that all
children of a spark plan
- // have same number of output partitions.
// Note that, `BroadcastQueryStageExec` is a valid case:
// If a join has been optimized from shuffled join to broadcast join, then
the one side is
// `BroadcastQueryStageExec` and other side is `ShuffleQueryStageExec`. It
can coalesce the
// shuffle side as we do not expect broadcast exchange has same partition
number.
+ case join: BroadcastHashJoinExec =>
join.children.flatMap(collectCoalesceGroups)
+ case join: BroadcastNestedLoopJoinExec =>
join.children.flatMap(collectCoalesceGroups)
+ // If not all leaf nodes are exchange query stages, it's not safe to
reduce the number of
+ // shuffle partitions, because we may break the assumption that all
children of a spark plan
+ // have same number of output partitions.
case p if p.collectLeaves().forall(_.isInstanceOf[ExchangeQueryStageExec])
=>
Review Comment:
> It seems we should improve this line to check if the child of
`BroadcastQueryStageExec` exists `Union`.
In fact, I have tested the queries with various join types and join types
with 'union', and found that `ShuffledHashJoin` and `SortMergeJoin` work
normally, while `BroadcastJoin` and `CartesianProduct` with `Union` fail. The
former case we should ignore broadcast joins, while the latter requires
additional detection. Maybe we should add the union check and solve the latter
cases in this pr?
--
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]
