You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@phoenix.apache.org by "Bin Shi (JIRA)" <ji...@apache.org> on 2018/09/25 17:51:00 UTC
[jira] [Commented] (PHOENIX-4926) Disentangle the guidepost data
from the actually materialized scans
[ https://issues.apache.org/jira/browse/PHOENIX-4926?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16627717#comment-16627717 ]
Bin Shi commented on PHOENIX-4926:
----------------------------------
[~lhofhansl],
Hi Lars, based on our internal discussions and the discussions on PHOENIX-4594, I opened this Jira.
> Disentangle the guidepost data from the actually materialized scans
> -------------------------------------------------------------------
>
> Key: PHOENIX-4926
> URL: https://issues.apache.org/jira/browse/PHOENIX-4926
> Project: Phoenix
> Issue Type: Improvement
> Reporter: Bin Shi
> Assignee: Bin Shi
> Priority: Major
>
> Currently, when we enable stats for intra-region read parallelization, the granularity of guide posts is the same as the granularity of parallel scans, which blocks us from using smaller guide posts in query complexity estimation and query optimization for better accuracy.
> We can define the target intra-region parallelism so that a scan is an integral multiple of the defined GUIDE_POSTS_WIDTH value and its value is between [1, the number of threads in thread pool or PRC handlers on Region server].
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)