[
https://issues.apache.org/jira/browse/HIVE-26986?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17936994#comment-17936994
]
Sungwoo Park edited comment on HIVE-26986 at 3/20/25 3:29 AM:
--------------------------------------------------------------
For the record, HIVE-26986 is actually a bug-fix, although it affects only a
limited class of queries (e.g., only query 71 is affected in the TPC-DS
benchmark).
As for hybridgrace_hashjoin_2 (and also query 23-a in the TPC-DS benchmark),
the fix accidentally applies shared-work-optimization to a part of the query
plan that is not visited by shared-work-optimization before HIVE-26986, thus
giving rise to unintended side-effects. It can be thought of as an example
where shared-work-optimization does not always produce better plans with
reduced execution time.
Hopefully HIVE-28833 will fix this issue.
was (Author: glapark):
For the record, HIVE-26986 is actually a bug-fix, although it affects only a
limited class of queries (e.g., only query 71 is affected in the TPC-DS
benchmark).
As for hybridgrace_hashjoin_2 (and also query 23-a in the TPC-DS benchmark),
the fix accidentally applies shared-work-optimization to a part of the query
plan that is not visited by shared-work-optimization before HIVE-26986, thus
giving rise to unintended side-effects. It can be thought of as an example
where shared-work-optimization does not always produce better plans with
reduced execution time.
Hopefully HIVE-28833 will this issue.
> ParallelEdgeFixer adds redundant reduce sink operators
> ------------------------------------------------------
>
> Key: HIVE-26986
> URL: https://issues.apache.org/jira/browse/HIVE-26986
> Project: Hive
> Issue Type: Sub-task
> Affects Versions: 4.0.0-alpha-2
> Reporter: Seonggon Namgung
> Assignee: Seonggon Namgung
> Priority: Major
> Labels: hive-4.1.0-must, pull-request-available
> Fix For: 4.1.0
>
> Attachments: Query71 OperatorGraph.png, Query71 TezDAG.png
>
> Time Spent: 50m
> Remaining Estimate: 0h
>
> A DAG created by OperatorGraph is not equal to the corresponding DAG that is
> submitted to Tez.
> Because of this problem, ParallelEdgeFixer reports a pair of normal edges to
> a parallel edge.
> We observe this problem by comparing OperatorGraph and Tez DAG when running
> TPC-DS query 71 on 1TB ORC format managed table.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
