Riza Suminto has posted comments on this change. ( http://gerrit.cloudera.org:8080/20009 )
Change subject: IMPALA-12183: Fix cardinality clamping across aggregation phases ...................................................................... Patch Set 5: (1 comment) http://gerrit.cloudera.org:8080/#/c/20009/5/testdata/workloads/functional-planner/queries/PlannerTest/tpcds/tpcds-q67.test File testdata/workloads/functional-planner/queries/PlannerTest/tpcds/tpcds-q67.test: http://gerrit.cloudera.org:8080/#/c/20009/5/testdata/workloads/functional-planner/queries/PlannerTest/tpcds/tpcds-q67.test@512 PS5, Line 512: mem > This is a significant reduction in memory estimate. Is this because the car This is because prior to this patch, each agg class is capped with inputCardinality = getChild(0).getCardinality(); In this case, inputCardinality = 15.09M, which is all rows coming out of 16:EXCHANGE. But according to my discussion with Aman, since this is FIRST_MERGE agg, not all 15.09M goes through all 9 agg classes. Instead, each agg stream will pass to their own agg class. For example, only 1 row will pass to Class 8 here, and the rest of (15.09M - 1) rows distributed to other 8 agg classes separately. -- To view, visit http://gerrit.cloudera.org:8080/20009 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I1d414fe56b027f887c7f901d8a6799a388b16b95 Gerrit-Change-Number: 20009 Gerrit-PatchSet: 5 Gerrit-Owner: Riza Suminto <[email protected]> Gerrit-Reviewer: Abhishek Rawat <[email protected]> Gerrit-Reviewer: Aman Sinha <[email protected]> Gerrit-Reviewer: Impala Public Jenkins <[email protected]> Gerrit-Reviewer: Kurt Deschler <[email protected]> Gerrit-Reviewer: Riza Suminto <[email protected]> Gerrit-Comment-Date: Fri, 09 Jun 2023 23:57:58 +0000 Gerrit-HasComments: Yes
