Tim Armstrong has posted comments on this change. ( http://gerrit.cloudera.org:8080/12535 )
Change subject: IMPALA-8014: Incorrect FK/PK cardinality estimation ...................................................................... Patch Set 4: (1 comment) http://gerrit.cloudera.org:8080/#/c/12535/4//COMMIT_MSG Commit Message: http://gerrit.cloudera.org:8080/#/c/12535/4//COMMIT_MSG@32 PS4, Line 32: Turns out that the change caused a spill test to fail. The new plan is > I agree that your solution is more stable. Fudging with magic numbers is as I would look at the plan that was generated before this change (join order and strategy), then modify the query by adding straight_join and BROADCAST/PARTITIONED hints to force that join order. Actually I guess if the join order didn't change, then it could be that the memory estimates were lower and it allocated smaller buffer sizes. In that case we could force the buffer sizes to be larger, but that seems overkill. -- To view, visit http://gerrit.cloudera.org:8080/12535 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: Id0db82564596b0a9271b89b8e3f7a3cf92d306da Gerrit-Change-Number: 12535 Gerrit-PatchSet: 4 Gerrit-Owner: Paul Rogers <[email protected]> Gerrit-Reviewer: Bharath Vissapragada <[email protected]> Gerrit-Reviewer: Impala Public Jenkins <[email protected]> Gerrit-Reviewer: Paul Rogers <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-Comment-Date: Mon, 04 Mar 2019 20:30:40 +0000 Gerrit-HasComments: Yes
