impala-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alex Behm (Code Review)" <>
Subject [Impala-ASF-CR] IMPALA-5856: Fix outer join predicate assignment.
Date Thu, 14 Sep 2017 00:00:26 GMT
Alex Behm has posted comments on this change.

Change subject: IMPALA-5856: Fix outer join predicate assignment.

Patch Set 1:

File fe/src/main/java/org/apache/impala/analysis/

PS1, Line 1381: // Right-most full-outer join table ref that is equal to or to the left of
              :     // table ref materializing 'tids'.
> Trying to visualize this hurts my brain :) Not sure if it helps to understa
Agree, difficult to comprehend. Removed.

PS1, Line 1385: this
> What is "this" refer to? After reading this more carefully, I think I under
Rephrased. The comment is supposed to describe what 'ojRef' is - that's why it's above 'ojRef'.

PS1, Line 1398: globalState_.ojClauseByConjunct.get(e.getId());
> getOjRef(e)?
File testdata/workloads/functional-planner/queries/PlannerTest/outer-joins.test:

PS1, Line 1020: from the On-clause of a left outer join
> Out of curiosity, did you check if we have the same issue if there is a lef
We have a test for that in L634. Please check if you think it's sufficient.

I don't think those cases are similar because if the FOJ comes after the LOJ, then only the
FOJ On-clause predicates are considered to be full-outer-joined (and will be handled by the
non-full-outer join assignment code path).

Line 1028:      and t1.bigint_col > 10 and t2.bigint_col > 30
> To make it more interesting, add a where clause with one predicate that ref
Good idea. Done.

To view, visit
To unsubscribe, visit

Gerrit-MessageType: comment
Gerrit-Change-Id: I93db34d988cb66e00aa05d7dc161e0ca47042acb
Gerrit-PatchSet: 1
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-Owner: Alex Behm <>
Gerrit-Reviewer: Alex Behm <>
Gerrit-Reviewer: Dimitris Tsirogiannis <>
Gerrit-HasComments: Yes

View raw message