[jira] [Commented] (KYLIN-2570) source code is that TopNCounter merge method problem

["Shaofeng SHI (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/KYLIN-2570?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16270218#comment-16270218 ] 

Shaofeng SHI commented on KYLIN-2570:
-------------------------------------

Hi, sorry for the late response; This is a good question; for the answer you can refer to this blog:
https://kylin.apache.org/blog/2016/03/19/approximate-topn-measure/

And the algorithm that it uses:
https://arxiv.org/pdf/1401.0702.pdf

Btw, the code has been improved for performance since what you saw, please check the latest version.

> source code is that TopNCounter merge method problem
> ----------------------------------------------------
>
>                 Key: KYLIN-2570
>                 URL: https://issues.apache.org/jira/browse/KYLIN-2570
>             Project: Kylin
>          Issue Type: Test
>            Reporter: aijuanbo
>
> In org.apache.kylin.measure.topn.TopNCounter have a merge method,
> I have a question that why notDuplicateItems offer(item, m2);
> I think is this.offer(item,entry.getValue().getValue().count);
> because item is not found in TopNCounter,it is first offer 



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)