impala-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tim Armstrong (Code Review)" <ger...@cloudera.org>
Subject [Impala-ASF-CR] IMPALA-5612: join inversion should factor in parallelism
Date Wed, 05 Jul 2017 21:57:18 GMT
Tim Armstrong has posted comments on this change.

Change subject: IMPALA-5612: join inversion should factor in parallelism
......................................................................


Patch Set 1:

(3 comments)

http://gerrit.cloudera.org:8080/#/c/7351/1/fe/src/main/java/org/apache/impala/planner/Planner.java
File fe/src/main/java/org/apache/impala/planner/Planner.java:

Line 425:         // Invert the join if doing so reduces the estimated time to execute the
join.
> Possibly move to a separate method?
Done


Line 434:         // * it costs the same to process each build and probe row.
> The hash table build is a two pass while the probe is one. 
Done


Line 472:             perRowCost = rhsCard * rhsAvgRowSize;
> Should check for overflows whenever multiplication is used.
It's floating point arithmetic so it will just overflow to "infinity", which is ok.


-- 
To view, visit http://gerrit.cloudera.org:8080/7351
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: Icacea4565ce25ef15aaab014684c9440dd501d4e
Gerrit-PatchSet: 1
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-Owner: Tim Armstrong <tarmstrong@cloudera.com>
Gerrit-Reviewer: Alex Behm <alex.behm@cloudera.com>
Gerrit-Reviewer: Mostafa Mokhtar <mmokhtar@cloudera.com>
Gerrit-Reviewer: Tim Armstrong <tarmstrong@cloudera.com>
Gerrit-HasComments: Yes

Mime
View raw message