Sailesh Mukil has posted comments on this change. Change subject: IMPALA-3804: Push per-split filtering into scanners ......................................................................
Patch Set 3: (1 comment) > Will respond to individual comments in next patch set, but want to > point out that filters can arrive at any time during a scan's > execution. So checking early and often is what we do for now - a > filter that may not have applied when issuing scan ranges may have > arrived and be applicable at the per-row-group scope. Ah, that makes sense. Do you think it's worth doing a multi-level check? i.e. once in IssueInitialRanges() and once in ProcessSplit()? Or will that not help us in the average case? http://gerrit.cloudera.org:8080/#/c/3561/3/be/src/exec/base-sequence-scanner.cc File be/src/exec/base-sequence-scanner.cc: PS3, Line 135: stream_->scan_range()->Cancel(Status::CANCELLED); As we spoke, this should be consistent with the other scanners, i.e. cancel "this" range in all scanners if necessary, or just return here like the parquet and text scanners do. -- To view, visit http://gerrit.cloudera.org:8080/3561 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I9f92178f642695e0e9ef901373a5e9f2878a78ce Gerrit-PatchSet: 3 Gerrit-Project: Impala Gerrit-Branch: cdh5-trunk Gerrit-Owner: Henry Robinson <[email protected]> Gerrit-Reviewer: Dan Hecht <[email protected]> Gerrit-Reviewer: Henry Robinson <[email protected]> Gerrit-Reviewer: Sailesh Mukil <[email protected]> Gerrit-HasComments: Yes
