impala-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Pranay Singh (Code Review)" <>
Subject [Impala-ASF-CR] IMPALA-5525 Extend TestScannersFuzzing to test uncompressed parquet
Date Tue, 26 Sep 2017 17:29:39 GMT
Pranay Singh has posted comments on this change. ( )

Change subject: IMPALA-5525 Extend TestScannersFuzzing to test uncompressed parquet

Patch Set 4:

File tests/query_test/
PS2, Line 97: if table_format.file_format != 'parquet':
> Another option is to keep the clone code in run_fuzz_test, but change the a
File tests/query_test/
PS4, Line 101:     """ Clone an existing parquet table with codec as none in the
             :         unique database. This cloned table is passed to run_fuzz_test
             :         which clones the table and corrupts the table. The test later
             :         checks that there is no crash while performing SQL queries on
             :         a corrupt table.
             :     """
> I think this comment should focus on why this test is different from the ot
PS4, Line 111:     db_name = unique_database
> I would prefer to emphasize that the source and destination are the unique_
PS4, Line 117: functional_parquet.alltypes
> Can we extend this to do fuzzing on decimal_tbl as well? I was thinking thi
PS4, Line 118: .format(fq_tbl_name))
> This indentation is a bit awkward. I don't think .format should be on its o
Moved the format in the same line as select.

To view, visit
To unsubscribe, visit

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I760de7203a51cf82b16016fa8043cadc7c8325bc
Gerrit-Change-Number: 8056
Gerrit-PatchSet: 4
Gerrit-Owner: Pranay Singh
Gerrit-Reviewer: Joe McDonnell <>
Gerrit-Reviewer: Pranay Singh
Gerrit-Comment-Date: Tue, 26 Sep 2017 17:29:39 +0000
Gerrit-HasComments: Yes

  • Unnamed multipart/alternative (inline, 8-Bit, 0 bytes)
View raw message