Tim Armstrong has posted comments on this change. Change subject: IMPALA-3311: fix string data coming out of aggs in subplans ......................................................................
Patch Set 2: (5 comments) http://gerrit.cloudera.org:8080/#/c/2929/2/be/src/exec/partitioned-aggregation-node.cc File be/src/exec/partitioned-aggregation-node.cc: Line 367: // String data produced by the aggregations is from local expr allocations, which will The comment is a little misleading: only strings allocated by Serialize()/Finalize() are local I think. The intermediate results of the agg should be non-local allocations. Maybe just clarify that it's "String data returned by Serialize() or Finalize()" Line 380: row_batch->MarkNeedToReturn(); I think we can break out of the loop early here, right, since MarkNeedToReturn() is idempotent? Line 387: RowBatch const? If there's some reason it can't be const, don't worry about it. http://gerrit.cloudera.org:8080/#/c/2929/2/testdata/workloads/targeted-perf/queries/subplan_aggregation.test File testdata/workloads/targeted-perf/queries/subplan_aggregation.test: Line 2: ---- QUERY: subplan_aggregation Maybe we should move this to perf-regression? It seems like a niche-enough case that we don't need to run it as part of the standard perf workload. I'm not sure how often/if we run perf-regression but it seems like the right place for it. Line 3: trival trivial -- To view, visit http://gerrit.cloudera.org:8080/2929 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Iada891504c261ba54f4eb8c9d7e4e5223668d7b9 Gerrit-PatchSet: 2 Gerrit-Project: Impala Gerrit-Branch: cdh5-trunk Gerrit-Owner: Skye Wanderman-Milne <[email protected]> Gerrit-Reviewer: Dan Hecht <[email protected]> Gerrit-Reviewer: Skye Wanderman-Milne <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-HasComments: Yes
