ulysses-you commented on a change in pull request #32391:
URL: https://github.com/apache/spark/pull/32391#discussion_r623574701
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/joins.scala
##########
@@ -376,7 +377,8 @@ trait JoinSelectionHelper {
* dynamic.
*/
private def canBuildLocalHashMapBySize(plan: LogicalPlan, conf: SQLConf):
Boolean = {
- plan.stats.sizeInBytes < conf.autoBroadcastJoinThreshold *
conf.numShufflePartitions
+ plan.stats.sizeInBytes <
+ autoBroadcastJoinThreshold(plan.stats.isAdaptive, conf) *
conf.numShufflePartitions
Review comment:
yeah, I have the similar thought. The current condition of converting
join to shuffled hash join is a bit rough that assume the data is not skewed
for all partition.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
