You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@kylin.apache.org by "Yaqian Zhang (Jira)" <ji...@apache.org> on 2020/01/02 07:45:00 UTC
[jira] [Commented] (KYLIN-4310) "select 1 from tb_name limit 1"
seems to scan all segments and dicts
[ https://issues.apache.org/jira/browse/KYLIN-4310?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17006642#comment-17006642 ]
Yaqian Zhang commented on KYLIN-4310:
-------------------------------------
Hi Congling:
I execute this simple sql and found that Kylin does take lots of resources and time for it.
Maybe it's a meaningful job to find out the reason and optimize it.
> "select 1 from tb_name limit 1" seems to scan all segments and dicts
> --------------------------------------------------------------------
>
> Key: KYLIN-4310
> URL: https://issues.apache.org/jira/browse/KYLIN-4310
> Project: Kylin
> Issue Type: Improvement
> Components: Query Engine
> Reporter: Congling Xia
> Priority: Major
>
> Hey! Some users of our kylin system tend to send SQL like
>
> {code:java}
> select 1 from <table_name> limit 1;
> {code}
> as a validate query. After checking the log, I found that Kylin seems to scans all segments of a cube and load all dictionaries of each columns. More HBase scans and HDFS I/O will take place in the process, and may lead to slow response.
> The dictionaries may cache in and out frequently in `DictionaryManager`. It is worse when dictionaries are big files and queries like that execute concurrently.
> It seems hard work for Kylin to execute that simple SQL...
> Any ideas for optimization?
>
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)