[
https://issues.apache.org/jira/browse/IMPALA-13902?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18038816#comment-18038816
]
ASF subversion and git services commented on IMPALA-13902:
----------------------------------------------------------
Commit bc997052520cd514543dc5b3630fedc2710d8dc2 in impala's branch
refs/heads/master from Steve Carlin
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=bc9970525 ]
IMPALA-13902: Calcite planner: Implement is_spool_query_results
The is_spool_query_results query option is now supported in Calcite. The
returnAtMostOneRow method is now implemented to support this.
PlanRootSink is refactored to extract sanitizing query options (a new
method sanitizeSpoolingOptions()) out of
PlanRootSink.computeResourceProfile(). The bulk of memory bounding
calculation is also extracted out to a new class SpoolingMemoryBound.
Added "sleep" in ImpalaOperatorTable.java since some EE tests related to
result spooling calls sleep() function. Changed ImpalaPlanRel to extends
RelNode interface.
A sanity test has been added to calcite.test, but the bulk of the
testing will be done through the Impala test framework when it is
enabled.
Testing:
- Pass FE tests PlannerTest#testResultSpooling, TpcdsCpuCostPlannerTest,
and all java tests under calcite-planner project.
- Pass query_test/test_result_spooling.py and
custom_cluster/test_result_spooling.py.
Co-authored-by: Riza Suminto
Change-Id: I5b9bf49e2874ee12de212b892bd898c296774c6f
Reviewed-on: http://gerrit.cloudera.org:8080/23562
Reviewed-by: Riza Suminto <[email protected]>
Reviewed-by: Impala Public Jenkins <[email protected]>
Tested-by: Impala Public Jenkins <[email protected]>
> returnsAtMostOneRow should be part of the planner interface
> -----------------------------------------------------------
>
> Key: IMPALA-13902
> URL: https://issues.apache.org/jira/browse/IMPALA-13902
> Project: IMPALA
> Issue Type: Sub-task
> Reporter: Steve Carlin
> Priority: Major
>
> In AnalysisContext, there is a method "returnsAtMostOneRow" on the SelectStmt
> that is used for result spooling.
> This is inaccessible in its current state by the Calcite planner since it is
> on StatementBase. The code should be refactored to handle this.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
