[
https://issues.apache.org/jira/browse/KYLIN-3883?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16800332#comment-16800332
]
Shaofeng SHI commented on KYLIN-3883:
-------------------------------------
It is different approach. In KYLIN-3360, it need the count column be a
dimension, which is costy in some cases. But this JIRA, it does't need the
coiunt column be on dimension; Kylin will pre-calcuate the "count(col)" value
during cube building phase.
> Kylin supports column count aggregation
> ---------------------------------------
>
> Key: KYLIN-3883
> URL: https://issues.apache.org/jira/browse/KYLIN-3883
> Project: Kylin
> Issue Type: New Feature
> Components: Query Engine
> Affects Versions: all
> Reporter: xiaodongzhang
> Assignee: xiaodongzhang
> Priority: Major
> Fix For: v3.0.0
>
>
> Kylin目前只支持对常量1进行count聚合预计算，即count(1)，count(col_1)的查询内部全部重写到对count(1)的查询，这样会导致一个问题：当col_1列中存在null值时，count(col_1)的查询结果并不准确。这样就导致Kylin的查询结果和Hive、Spark等不一致。该Patch中，提供了对count(col_1)的支持。
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
