[
https://issues.apache.org/jira/browse/PHOENIX-4568?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Csaba Skrabak updated PHOENIX-4568:
-----------------------------------
Description:
AggregateIT.testTrimDistinct case is introduced in the fix of PHOENIX-4139.
Trace-debugging the test reveals that the GroupBy class may store duplicates of
accessor objects in its list fields, keyExpressions and expressions.
"Since the second trim expression is the same a the first one, the group by
(district turns into a group by) of the second one should be ignored as it
serves no purpose. That is what occurs when you do a select without the
distinct. Perhaps this logic is missing from GroupByCompiler?"
https://issues.apache.org/jira/browse/PHOENIX-4139?focusedCommentId=16274531&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-16274531
I have not yet found any test case in which this internal behavior would cause
an error but still.
was:
AggregateIT.testTrimDistinct case is introduced in the fix of PHOENIX-4139.
Trace-debugging the test reveals that the GroupBy class may store duplicates of
accessor objects in its list fields, keyExpressions and expressions.
"Since the second trim expression is the same a the first one, the group by
(district turns into a group by) of the second one should be ignored as it
serves no purpose. That is what occurs when you do a select without the
distinct. Perhaps this logic is missing from GroupByCompiler?"
https://issues.apache.org/jira/browse/PHOENIX-4139?focusedCommentId=16274531&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-16274531
> Duplicate entries in the GroupBy structure when running
> AggregateIT.testTrimDistinct
> ------------------------------------------------------------------------------------
>
> Key: PHOENIX-4568
> URL: https://issues.apache.org/jira/browse/PHOENIX-4568
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 4.12.0
> Environment: minicluster
> Reporter: Csaba Skrabak
> Assignee: Csaba Skrabak
> Priority: Minor
> Fix For: 4.14.0
>
>
> AggregateIT.testTrimDistinct case is introduced in the fix of PHOENIX-4139.
> Trace-debugging the test reveals that the GroupBy class may store duplicates
> of accessor objects in its list fields, keyExpressions and expressions.
> "Since the second trim expression is the same a the first one, the group by
> (district turns into a group by) of the second one should be ignored as it
> serves no purpose. That is what occurs when you do a select without the
> distinct. Perhaps this logic is missing from GroupByCompiler?"
> https://issues.apache.org/jira/browse/PHOENIX-4139?focusedCommentId=16274531&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-16274531
> I have not yet found any test case in which this internal behavior would
> cause an error but still.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
