You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hive.apache.org by Ashutosh Chauhan <ha...@apache.org> on 2017/07/08 00:41:54 UTC
Re: Review Request 60349: Scalar sub-query: sq_count_check not
required if gby keys are constant
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/60349/#review179973
-----------------------------------------------------------
ql/src/java/org/apache/hadoop/hive/ql/optimizer/calcite/rules/HiveRemoveSqCountCheck.java
Lines 42 (patched)
<https://reviews.apache.org/r/60349/#comment254951>
Add some comment on what rule does.
ql/src/java/org/apache/hadoop/hive/ql/optimizer/calcite/rules/HiveRemoveSqCountCheck.java
Lines 74 (patched)
<https://reviews.apache.org/r/60349/#comment254952>
Comment seems misplaced.
ql/src/java/org/apache/hadoop/hive/ql/optimizer/calcite/rules/HiveSubQueryRemoveRule.java
Lines 178 (patched)
<https://reviews.apache.org/r/60349/#comment254954>
Lets add a flag (with default off) hive.optimize.remove.sq_count_check which doesnt add this project, that way we avoid problem of both join order as well as join merging.
ql/src/java/org/apache/hadoop/hive/ql/optimizer/calcite/rules/HiveSubQueryRemoveRule.java
Lines 202 (patched)
<https://reviews.apache.org/r/60349/#comment254953>
can be deleted.
ql/src/java/org/apache/hadoop/hive/ql/parse/CalcitePlanner.java
Lines 1533-1536 (patched)
<https://reviews.apache.org/r/60349/#comment254955>
Rule will fire only when flag is on.
- Ashutosh Chauhan
On June 26, 2017, 9:06 p.m., Vineet Garg wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/60349/
> -----------------------------------------------------------
>
> (Updated June 26, 2017, 9:06 p.m.)
>
>
> Review request for hive and Ashutosh Chauhan.
>
>
> Bugs: HIVE-16793
> https://issues.apache.org/jira/browse/HIVE-16793
>
>
> Repository: hive-git
>
>
> Description
> -------
>
> For scalar subqueries we skip adding UDF sq_count_check which have aggregates but we add it if there is a explicit group by present. We can skip it if group by is present but group by keys are constant. This patch adds a rule to do this optimization.
>
>
> Diffs
> -----
>
> ql/src/java/org/apache/hadoop/hive/ql/optimizer/calcite/rules/HiveRemoveSqCountCheck.java PRE-CREATION
> ql/src/java/org/apache/hadoop/hive/ql/optimizer/calcite/rules/HiveSubQueryRemoveRule.java 83d3f7436d
> ql/src/java/org/apache/hadoop/hive/ql/parse/CalcitePlanner.java 88054e7379
> ql/src/test/results/clientnegative/subquery_scalar_multi_rows.q.out 0a780db7ef
> ql/src/test/results/clientpositive/llap/subquery_scalar.q.out ceccc88b2e
> ql/src/test/results/clientpositive/llap/subquery_select.q.out 440e9d7b55
> ql/src/test/results/clientpositive/perf/query23.q.out ebd2271108
> ql/src/test/results/clientpositive/perf/query44.q.out 07187b14d4
> ql/src/test/results/clientpositive/perf/query54.q.out b9d0b8b376
> ql/src/test/results/clientpositive/perf/query58.q.out 5b37f36386
> ql/src/test/results/clientpositive/perf/query6.q.out 03fff79951
>
>
> Diff: https://reviews.apache.org/r/60349/diff/2/
>
>
> Testing
> -------
>
> -Pre-commit testing
> - new tests
>
>
> Thanks,
>
> Vineet Garg
>
>