Aman Sinha has posted comments on this change. ( http://gerrit.cloudera.org:8080/16242 )
Change subject: IMPALA-9979: part 2: partitioned top-n ...................................................................... Patch Set 30: (5 comments) http://gerrit.cloudera.org:8080/#/c/16242/30/testdata/workloads/functional-planner/queries/PlannerTest/analytic-rank-pushdown.test File testdata/workloads/functional-planner/queries/PlannerTest/analytic-rank-pushdown.test: http://gerrit.cloudera.org:8080/#/c/16242/30/testdata/workloads/functional-planner/queries/PlannerTest/analytic-rank-pushdown.test@55 PS30, Line 55: rnk nit: This extra rnk is not needed since doing a select * from the subquery should also fetch this (same for other queries in this file). http://gerrit.cloudera.org:8080/#/c/16242/30/testdata/workloads/functional-planner/queries/PlannerTest/analytic-rank-pushdown.test@755 PS30, Line 755: # Analytic predicate can be migrated into inline view, but not through the analytic Is this true even if the rank predicate is on the top level subquery, not inner subquery ? I suppose in this case it is because of the conjuncts ..one of which is on top level subquery and another on inner. http://gerrit.cloudera.org:8080/#/c/16242/30/testdata/workloads/functional-planner/queries/PlannerTest/analytic-rank-pushdown.test@1741 PS30, Line 1741: # Other predicates referencing rank() and row_number() can't be For the cases where the optimization is not applied maybe the DISTRIBUTEDPLAN section can be skipped ? http://gerrit.cloudera.org:8080/#/c/16242/30/testdata/workloads/functional-planner/queries/PlannerTest/limit-pushdown-partitioned-top-n.test File testdata/workloads/functional-planner/queries/PlannerTest/limit-pushdown-partitioned-top-n.test: http://gerrit.cloudera.org:8080/#/c/16242/30/testdata/workloads/functional-planner/queries/PlannerTest/limit-pushdown-partitioned-top-n.test@76 PS30, Line 76: # row_number() predicate on equality instead of range. In the previous patch (IMPALA-10296) equality predicate was disallowed except for limit 1. I might have missed the code change in this patch that allows this to happen. http://gerrit.cloudera.org:8080/#/c/16242/30/testdata/workloads/functional-planner/queries/PlannerTest/limit-pushdown-partitioned-top-n.test@777 PS30, Line 777: # Limit pushdown should not be applied. But partitioned top-n is created. It would be good to clarify if (a) no optimization was done (so the original Analytic Sort remains) (b) partitioned top-n with/without limit pushdown was applied. -- To view, visit http://gerrit.cloudera.org:8080/16242 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: Ic638af9495981d889a4cb7455a71e8be0eb1a8e5 Gerrit-Change-Number: 16242 Gerrit-PatchSet: 30 Gerrit-Owner: Tim Armstrong <[email protected]> Gerrit-Reviewer: Aman Sinha <[email protected]> Gerrit-Reviewer: David Rorke <[email protected]> Gerrit-Reviewer: Impala Public Jenkins <[email protected]> Gerrit-Reviewer: Qifan Chen <[email protected]> Gerrit-Reviewer: Shant Hovsepian <[email protected]> Gerrit-Reviewer: Thomas Tauber-Marshall <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-Comment-Date: Mon, 08 Feb 2021 19:05:57 +0000 Gerrit-HasComments: Yes
