Thomas Marshall has posted comments on this change. ( http://gerrit.cloudera.org:8080/10394 )
Change subject: IMPALA-110 (part 2): Refactor PartitionedAggregationNode ...................................................................... Patch Set 6: (2 comments) http://gerrit.cloudera.org:8080/#/c/10394/6/be/src/exec/aggregator.h File be/src/exec/aggregator.h: http://gerrit.cloudera.org:8080/#/c/10394/6/be/src/exec/aggregator.h@26 PS6, Line 26: #include "runtime/mem-tracker.h" > Nit: Can we forward-declare MemTracker instead of pulling in the whole head No, unique_ptr requires a full definition to determine the size http://gerrit.cloudera.org:8080/#/c/10394/4/be/src/exec/grouping-aggregator.cc File be/src/exec/grouping-aggregator.cc: http://gerrit.cloudera.org:8080/#/c/10394/4/be/src/exec/grouping-aggregator.cc@91 PS4, Line 91: is_streaming_preagg_(tnode.agg_node.use_streaming_preaggregation), > Right, I was imaging that eventually we'll have separate thrift nodes for A I'm not sure having a TAggregationNode and a TStreamingAggregationNode is the right way to go, since they would have identical fields. And of course we already handle other nodes this way, eg. there's a TSortNode that could produce either a SortNode, PartialSortNode, or TopNNode depending on a flag in the TSortNode -- To view, visit http://gerrit.cloudera.org:8080/10394 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I9e7bb583f54aa4add3738bde7f57cf3511ac567e Gerrit-Change-Number: 10394 Gerrit-PatchSet: 6 Gerrit-Owner: Thomas Marshall <thomasmarsh...@cmu.edu> Gerrit-Reviewer: Alex Behm <alex.b...@cloudera.com> Gerrit-Reviewer: Dan Hecht <dhe...@cloudera.com> Gerrit-Reviewer: Thomas Marshall <thomasmarsh...@cmu.edu> Gerrit-Reviewer: Tim Armstrong <tarmstr...@cloudera.com> Gerrit-Reviewer: Vuk Ercegovac <vercego...@cloudera.com> Gerrit-Comment-Date: Thu, 28 Jun 2018 23:26:11 +0000 Gerrit-HasComments: Yes