Tim Armstrong has posted comments on this change. Change subject: IMPALA-2502: don't redundantly repartition grouping aggregations ......................................................................
Patch Set 7: (1 comment) http://gerrit.cloudera.org:8080/#/c/2414/7/testdata/workloads/functional-planner/queries/PlannerTest/aggregation.test File testdata/workloads/functional-planner/queries/PlannerTest/aggregation.test: Line 1066: select straight_join coalesce(l_orderkey, 123), count(*) > Can you explain why we still have an exchange in both example? Just trying I added some logging. In the second query it's comparing: SlotRef{type=BIGINT, id=2} vs FunctionCallExpr{name=coalesce, isStar=false, isDistinct=false, SlotRef{path=l_orderkey, type=BIGINT, id=0} NumericLiteral{value=123, type=BIGINT}} explain select straight_join ifnull(o_orderkey, 20), count(*) from tpch_parquet.lineitem join [shuffle] tpch_parquet.orders on coalesce(l_orderkey,123) = ifnull(o_orderkey, 20) group by 1; I guess it's not inferring that the output slot corresponding to the grouping expr of the aggregation and the function call in the join are in the same equivalence class? -- To view, visit http://gerrit.cloudera.org:8080/2414 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Iffdcfd3629b8a69bd23915e1adba3b8323cbbaef Gerrit-PatchSet: 7 Gerrit-Project: Impala Gerrit-Branch: cdh5-trunk Gerrit-Owner: Tim Armstrong <[email protected]> Gerrit-Reviewer: Alex Behm <[email protected]> Gerrit-Reviewer: Marcel Kornacker <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-HasComments: Yes
