Matthew Jacobs has posted comments on this change. Change subject: IMPALA-4134,IMPALA-3704: Kudu INSERT improvements ......................................................................
Patch Set 6: (1 comment) http://gerrit.cloudera.org:8080/#/c/4728/6/be/src/exec/kudu-table-sink.cc File be/src/exec/kudu-table-sink.cc: PS6, Line 330: kudu_error_counter_->value() > Good point, but this kinda means that IGNORE is buggy (at least theoretical True. I think we just need to make sure the Kudu folks don't change the behavior without telling us so we can work with them to make sure that won't happen in practice, e.g. I could imagine a few things to consider if they want to set error buffer limits so we can make sure the buffer is big enough to handle at least all conflicts. Or they offer a way to report errors different for different types. Ideally we could have a test case that ensures we don't hit any unexpected code paths, I'm not sure we can guarantee any test would overflow errors if they change the behavior in the future. I do have a test case that will result in key conflicts for >7000 rows, but depending on how they could change the code that might or might not result in errors overflowing. I'll start a conversation with them about this. -- To view, visit http://gerrit.cloudera.org:8080/4728 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I5542b9a061b01c543a139e8722560b1365f06595 Gerrit-PatchSet: 6 Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-Owner: Matthew Jacobs <m...@cloudera.com> Gerrit-Reviewer: Alex Behm <alex.b...@cloudera.com> Gerrit-Reviewer: Dan Hecht <dhe...@cloudera.com> Gerrit-Reviewer: Matthew Jacobs <m...@cloudera.com> Gerrit-Reviewer: Todd Lipcon <t...@apache.org> Gerrit-HasComments: Yes