[
https://issues.apache.org/jira/browse/PHOENIX-4807?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16539046#comment-16539046
]
Josh Elser edited comment on PHOENIX-4807 at 7/10/18 6:09 PM:
--------------------------------------------------------------
[~sergey.soldatov] was the brains behind getting to the bottom of this one.
Could you bring some of your findings (specifically, the examples on how stats
are used by a query) into the website, please?
was (Author: elserj):
[~sergey.soldatov] was the brains behind getting to the bottom of this one.
> Document query behavior when statistics unavailable for some table regions.
> ---------------------------------------------------------------------------
>
> Key: PHOENIX-4807
> URL: https://issues.apache.org/jira/browse/PHOENIX-4807
> Project: Phoenix
> Issue Type: Task
> Reporter: Guru Prateek Pinnadhari
> Assignee: Sergey Soldatov
> Priority: Major
>
> When troubleshooting an issue with retrieval of duplicate records for a
> select query on same PK, the behavior seen was: query plan's collection of
> Scan objects were such that they led to generation of overlapping scans on a
> salted table's regions.
> A contributing factor was statistics being unavailable for the terminal
> regions of the table. Filing Jira to add a note about this on
> [https://phoenix.apache.org/update_statistics.html]
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
