[GitHub] [spark] cloud-fan opened a new pull request #24336: [SPARK-27430][SQL] BroadcastNestedLoopJoinExec should support all join types

[GitBox <gi...@apache.org>]

cloud-fan opened a new pull request #24336: [SPARK-27430][SQL] BroadcastNestedLoopJoinExec should support all join types
URL: https://github.com/apache/spark/pull/24336
 
 
   ## What changes were proposed in this pull request?
   
   `BroadcastHashJoinExec` can't build left side for left join, because the build side will become a hash relation, and hash join does not iterate this hash relation. That's why `BroadcastHashJoinExec` does not support full outer join.
   
   `BroadcastNestedLoopJoinExec` doesn't have this limitation. It iterates both join sides, and can build any side no matter what the join type is. That's why `BroadcastNestedLoopJoinExec` is the last choice in join selection.
   
   This PR fixes the join selection, to not be constrained by the join type when picking the build side for `BroadcastNestedLoopJoinExec`.
   
   ## How was this patch tested?
   
   updated tests in `BroadcastJoinSuite`

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org