[
https://issues.apache.org/jira/browse/PHOENIX-1006?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14013270#comment-14013270
]
jay wong commented on PHOENIX-1006:
-----------------------------------
[~jamestaylor]
yep. I saw it sort per region. but it sout one by one. not Parallel.
so. data increment. and region increment .as it is Parallel.
the performance descend huge
> 8x Performance enhancements in my group by query case.
> ------------------------------------------------------
>
> Key: PHOENIX-1006
> URL: https://issues.apache.org/jira/browse/PHOENIX-1006
> Project: Phoenix
> Issue Type: Improvement
> Affects Versions: 3.0.0
> Reporter: jay wong
> Attachments: PHOENIX-1006.patch, PHOENIX-1006v2.patch
>
>
> As a query sql is :
> SELECT count(1) as count,SUM(int_column) as sum_column, MAX(int_column) as
> max_column2,MIN(int_column) as min_column,AVG(int_column) as avg_column FROM
> table1 WHERE int_column IS NOT NULL GROUP BY int_column2 ORDER BY int_column
> DESC LIMIT 200;
> NOTE: The result has 55000 groups.
> AggregatePlan tell me when match group by. the resultIterator will be set as
> MergeSortRowKeyResultIterator.
> And the MergeSortRowKeyResultIterator need a OrderedResultIterator.
> So when you run a group by query with order or not . It ALWAYS has a ordered
> First.
> I think it's unnecessary at all.
> So I get all result from RegionServer and Aggregate by key.
> with the above sql. the execute RT is from 30s enhancements to 4s.
> ==============================================
> But I found If group by without order by. the result sequence could not be
> safeguarded.
> So the result sequence I sort by TreeSet with key.
> Now the result is same as query with MergeSortRowKeyResultIterator
--
This message was sent by Atlassian JIRA
(v6.2#6252)
