[ 
https://issues.apache.org/jira/browse/IMPALA-15195?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18108874#comment-18108874
 ] 

ASF subversion and git services commented on IMPALA-15195:
----------------------------------------------------------

Commit 8ffcc92013dacc74b85eb0bfa62c9475dc0060b4 in impala's branch 
refs/heads/master from Steve Carlin
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=8ffcc9201 ]

IMPALA-15261: Calcite Planner: Fix broken tests due to plan changes.

The commit IMPALA-15195 caused some significant plan changes for the
Calcite planner. As a result, some tests were broken when running
with the Calcite planner on.

In mt-dop-parquet-scheduling.test: Some node numbers were changed

In test_processing_cost.py: Plan changes caused this to fail. Since
this is a backend test, these failures are ok, but should be fixed
at some point. IMPALA-15260 has been filed for this.

In test_executor_groups: This is also a backend test. A Jira was
already filed for the memory estimate issue, IMPALA-14816. A comment
was added to this Jira to handle this problem too.

In test_iceberg.py: Fixed this by running with "straight_join", since
this shouldn't affect the integrity of the test.

Change-Id: I77683c0ab2f506b6c638255b63bde243b33dd8fb
Reviewed-on: http://gerrit.cloudera.org:8080/24691
Reviewed-by: Impala Public Jenkins <[email protected]>
Tested-by: Impala Public Jenkins <[email protected]>


> Calcite Planner: general performance improvements for tpcds
> -----------------------------------------------------------
>
>                 Key: IMPALA-15195
>                 URL: https://issues.apache.org/jira/browse/IMPALA-15195
>             Project: IMPALA
>          Issue Type: Sub-task
>            Reporter: Steve Carlin
>            Priority: Major
>
> Some of the tpcds queries are slower on the Calcite planner when compared 
> with the original planner.  Investigation needs to be done for these slower 
> queries.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to