ngsg opened a new pull request, #4406:
URL: https://github.com/apache/hive/pull/4406
<!--
Thanks for sending a pull request! Here are some tips for you:
1. If this is your first time, please read our contributor guidelines:
https://cwiki.apache.org/confluence/display/Hive/HowToContribute
2. Ensure that you have created an issue on the Hive project JIRA:
https://issues.apache.org/jira/projects/HIVE/summary
3. Ensure you have added or run the appropriate tests for your PR:
4. If the PR is unfinished, add '[WIP]' in your PR title, e.g.,
'[WIP]HIVE-XXXXX: Your PR title ...'.
5. Be sure to keep the PR description updated to reflect all changes.
6. Please write your PR title to summarize what this PR proposes.
7. If possible, provide a concise example to reproduce the issue for a
faster review.
-->
### What changes were proposed in this pull request?
<!--
Please clarify what changes you are proposing. The purpose of this section
is to outline the changes and how this PR fixes the issue.
If possible, please consider writing useful notes for better and faster
reviews in your PR. See the examples below.
1. If you refactor some codes with changing classes, showing the class
hierarchy will help reviewers.
2. If you fix some SQL features, you can provide some references of other
DBMSes.
3. If there is design documentation, please add the link.
4. If there is a discussion in the mailing list, please add the link.
-->
When setting the output name of ReduceSinkOperator, check whether the output
ReduceWork is contained in any MergeJoinWork's merged work list.
If so, set the output name to the name of the main work of that
MergeJoinWork.
### Why are the changes needed?
<!--
Please clarify why the changes are needed. For instance,
1. If you propose a new API, clarify the use case for a new API.
2. If you fix a bug, you can clarify why it is a bug.
-->
In short, ReduceSinkOperator may point a merged BaseWork of MergeJoinWork
instead of its main BaseWork.
When this happens, ReduceSinkOperator does not send intermediate data to
next Vertex in runtime.
For more details, let's assume the following partial OperatorGraph.
...
TS1--UNION1-RS1-GBY-DUMMYSTORE-MERGEJOIN
TS2-/
...
TezCompiler creates 5 BaseWorks for the above operators:
- MapWork1(TS1-RS1)
- MapWork2(TS2-(cloned)RS1)
- ReduceWork1(GBY-DUMMYSTORE)
- ReduceWork2(MERGEJOIN)
- MergeJoinWork1(ReduceWork1, ReduceWork2).
The output name of RS1 should be the name of MergeJoinWork1 because
ReduceWork1 will not be converted to a regular Tez Vertex. (cf. The name of
MergeJoinWork is equal to the name of its main work.)
Suppose that TezCompiler visits the operators in the following order: TS1,
UNION1, RS1, GBY, DUMMYSTORE, MERGEJOIN, TS2, UNION1, RS1.
TezCompiler creates ReduceWork1 when it visits DUMMYSTORE (after TS1,
UNION1, and RS1). At this point, the output name of RS1 is set to ReduceWork1.
TezCompiler cuts the parent-child relationship between RS1 and GBY and proceeds
to next operator, MERGEJOIN.
TezCompiler visits MERGEJOIN and indicates that ReduceWork1 is a merged work
of MergeJoinWork1. Therefore, TezCompiler changes the output name of RS1 to the
name of MergeJoinWork1. At this point, RS1's output name is ReduceWork2.
TezCompiler visits RS1 again through TS2-UNION1. It creates MapWork2 for
TS2-UNION1-RS1 and then it indicates that there is a following BaseWork for
RS1. When there is a following BaseWork after a ReduceSink, TezCompiler sets
the name of the following BaseWork to the ReduceSink's output name. At this
point, RS1's output name is set to ReduceWork1 again.
The problem happens here. Since RS1 and GBY do not have parent-child
relationship anymore, TezCompiler skips visiting MERGEJOIN again. So RS1's
output name will not be corrected in the remaining traversal, which leads to
HIVE-27303 issue.
### Does this PR introduce _any_ user-facing change?
<!--
Note that it means *any* user-facing change including all aspects such as
the documentation fix.
If yes, please clarify the previous behavior and the change this PR proposes
- provide the console output, description, screenshot and/or a reproducable
example to show the behavior difference if possible.
If possible, please also clarify if this is a user-facing change compared to
the released Hive versions or within the unreleased branches such as master.
If no, write 'No'.
-->
No
### How was this patch tested?
<!--
If tests were added, say they were added here. Please make sure to add some
test cases that check the changes thoroughly including negative and positive
cases if possible.
If it was tested in a way different from regular unit tests, please clarify
how you tested step by step, ideally copy and paste-able, so that other
reviewers can test and check, and descendants can verify in the future.
If tests were not added, please describe why they were not added and/or why
it was difficult to add.
-->
I added a new qfile test that runs the query explained in HIVE-27303 JIRA
page.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
