Todd Lipcon has posted comments on this change. ( http://gerrit.cloudera.org:8080/11297 )
Change subject: IMPALA-7477: Batch-oriented query set construction ...................................................................... Patch Set 4: (1 comment) http://gerrit.cloudera.org:8080/#/c/11297/4/be/src/service/hs2-util.cc File be/src/service/hs2-util.cc: http://gerrit.cloudera.org:8080/#/c/11297/4/be/src/service/hs2-util.cc@128 PS4, Line 128: ReserveSpace Just passing through here... is it possible that this reserving design will end up with an "accidentally quadratic" scenario in the case that rows are added a few at a time? I'm not sure at what level RowBatches get aggregated together into batches that get delivered to the client, but if you had a RowBatch size of 1000 rows, and some selective predicate such that each RowBatch only has one or two matches, would we end up calling AddRows() with num_rows=1 in a loop anywhere? If that's the case we may want to do something here like reserve the next larger power-of-two capacity rather than "size to fit this next batch" -- To view, visit http://gerrit.cloudera.org:8080/11297 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: Ibc87a84c34935d0d5841c7f5528eb802527fa809 Gerrit-Change-Number: 11297 Gerrit-PatchSet: 4 Gerrit-Owner: Tim Armstrong <tarmstr...@cloudera.com> Gerrit-Reviewer: Impala Public Jenkins <impala-public-jenk...@cloudera.com> Gerrit-Reviewer: Todd Lipcon <t...@apache.org> Gerrit-Reviewer: Zoltan Borok-Nagy <borokna...@cloudera.com> Gerrit-Comment-Date: Fri, 24 Aug 2018 22:58:39 +0000 Gerrit-HasComments: Yes
