impala-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tim Armstrong (Code Review)" <>
Subject [Impala-ASF-CR] IMPALA-3804: Re-enable per-scan filtering for sequence-based scanners
Date Fri, 01 Dec 2017 19:37:51 GMT
Tim Armstrong has posted comments on this change. ( )

Change subject: IMPALA-3804: Re-enable per-scan filtering for sequence-based scanners

Patch Set 1:

File be/src/exec/
PS2, Line 691: 
nit: i instead of j since we unnested the loop
File be/src/exec/
PS1, Line 497: 
> Currently I added the header checking functionality to BaseSequenceScanner,
The approaches seem pretty comparable in readability. I guess the major advantage of putting
it in ScanRangeMetadata is that we don't need to instantiate the scanner before checking the
filters, which could be moderately expensive (although not a show stopper). So I think I have
a slight preference for that approach.
File be/src/exec/
PS2, Line 512:       runtime_state_, this, partition, scan_range, filter_ctxs, expr_results_pool);
nit: we could probably just put this expression inline in the if() right? Not sure that pulling
it out makes it more readable.
PS2, Line 518: 
nit: bs_scanner != nullptr, just to be explicit.
PS2, Line 519:     if (VLOG_QUERY_IS_ON) {
nit: unnecessary vertical whitespace

To view, visit
To unsubscribe, visit

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I4b38c26bcbe67f83efcc65a1723d766626ae3d3e
Gerrit-Change-Number: 8684
Gerrit-PatchSet: 1
Gerrit-Owner: Zoltan Borok-Nagy <>
Gerrit-Reviewer: Dan Hecht <>
Gerrit-Reviewer: Henry Robinson <>
Gerrit-Reviewer: Tim Armstrong <>
Gerrit-Reviewer: Zoltan Borok-Nagy <>
Gerrit-Comment-Date: Fri, 01 Dec 2017 19:37:51 +0000
Gerrit-HasComments: Yes

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