ilooner commented on a change in pull request #1373: DRILL-6517: Hash-Join: If
not OK, exit early from prefetchFirstBatchFromBothSides
URL: https://github.com/apache/drill/pull/1373#discussion_r201807800
##########
File path:
exec/java-exec/src/main/java/org/apache/drill/exec/physical/impl/join/HashJoinBatch.java
##########
@@ -258,6 +251,13 @@ protected boolean prefetchFirstBatchFromBothSides() {
return false;
Review comment:
@Ben-Zvi I agree the case where only one side receives NONE is legit. I
think the issue @sohami was referring to is that **batchMemoryManager.update**
will be called for both sides even when only one side has NONE. Looking at the
javadoc for IterOutcome.NONE there is no guarantee that we will have a
container with a value record count set on it. Also looking at other operators
such as project, they do not seem to set the record count in the case where
they process an empty table, so we cannot assume that a container will have a
record count of 0 when receiving NONE.
Could you add a check to call update and debug statements for the left and
right side only if their IterOutcome is not NONE?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on 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
