[
https://issues.apache.org/jira/browse/PHOENIX-5937?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17147724#comment-17147724
]
Richard Antal commented on PHOENIX-5937:
----------------------------------------
Thank you [~stoty] for for checking the patch. I have uploaded v4 right now, it
fixes 1. 2.
I compared the following queries,
{code:java}
select * from SYSTEM.CATALOG where TABLE_NAME = 'US_POPULATION' order by
ORDINAL_POSITION;
select COLUMN_NAME from SYSTEM.CATALOG where TABLE_NAME = 'US_POPULATION' order
by ORDINAL_POSITION;{code}
And I found some difference in scan plan, debugged where the source of this
difference is.
In the BaseResultIterators line number 246-251 we add these empty key values to
the scan, but it only happened in case of !projector.projectEverything() was
true. (so the 2nd line)
Basically I just added an extra if statement for the given circumstances and
copied the lines which modified the scan plan.
> Order by on nullable column sometimes filters rows
> --------------------------------------------------
>
> Key: PHOENIX-5937
> URL: https://issues.apache.org/jira/browse/PHOENIX-5937
> Project: Phoenix
> Issue Type: Bug
> Components: core
> Affects Versions: 4.x, master
> Reporter: Istvan Toth
> Assignee: Richard Antal
> Priority: Major
> Attachments: PHOENIX-5937.master.v1.patch,
> PHOENIX-5937.master.v2.patch, PHOENIX-5937.master.v3.patch,
> PHOENIX-5937.master.v4.patch
>
>
> When selecting using order by on a nullable column, the rows where the order
> column is null are sometimes not returned.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
