Tim Armstrong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/9556 )

Change subject: IMPALA-6589: remove invalid DCHECK in parquet reader
......................................................................


Patch Set 1:

(3 comments)

http://gerrit.cloudera.org:8080/#/c/9556/1/be/src/exec/parquet-column-readers.cc
File be/src/exec/parquet-column-readers.cc:

http://gerrit.cloudera.org:8080/#/c/9556/1/be/src/exec/parquet-column-readers.cc@a175
PS1, Line 175:
> Nit: good to check that num_cached_levels GE 0
Done


http://gerrit.cloudera.org:8080/#/c/9556/1/be/src/exec/parquet-column-readers.cc@a175
PS1, Line 175:
> My guess is that this can occur if a corrupt Parquet file contains a litera
Csaba does raise an interesting point - the Parquet RLE spec may allow 0-length 
literal runs, which our code might treat as meaning "end of input". I guess 
that the spec didn't intend to allow this.

I could trigger this DCHECK without exercising that particular edge case, but 
it seems like it might be worth pursuing that line of thought further.


http://gerrit.cloudera.org:8080/#/c/9556/1/be/src/exec/parquet-column-readers.cc@a175
PS1, Line 175:
> It is feasible to add a test file that would have hit this and now is corre
Done



--
To view, visit http://gerrit.cloudera.org:8080/9556
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: Idd6e09f8c8cca8991be5b5b379f6420adaa97daa
Gerrit-Change-Number: 9556
Gerrit-PatchSet: 1
Gerrit-Owner: Tim Armstrong <[email protected]>
Gerrit-Reviewer: Csaba Ringhofer <[email protected]>
Gerrit-Reviewer: Lars Volker <[email protected]>
Gerrit-Reviewer: Pranay Singh
Gerrit-Reviewer: Tim Armstrong <[email protected]>
Gerrit-Comment-Date: Thu, 08 Mar 2018 22:30:40 +0000
Gerrit-HasComments: Yes

Reply via email to