[
https://issues.apache.org/jira/browse/KYLIN-3322?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16776848#comment-16776848
]
KANG-SEN LU commented on KYLIN-3322:
------------------------------------
Hi, Shaofeng:
Thanks for your response. I have two points to add.
# What if I already put SUM(X) in a separated cube, why do I have to add
SUM(X) into second cube while I am defining TOPN(X) in the second cube. If it
is just redundant metadata data, I will not complain about the extra human
effort. I am worried if the kylin may not be able to find the right cube to
compute SUM(X), because now there are two cubes both are, supposed, equally
qualified to answer the query. It will create more challenge to the cost
evaluation function to kylin.
# My experiment seems to suggest that when SUM(X) not group by B was issued,
the cost evaluation function sent the query to the cube containing both
TOPN(SUM(X)) and SUM(X) and, more importantly, it goes after TOPN(SUM(X)), then
perform SUM(X), that takes more than 20 seconds in my test case. If it goes
after SUM(X) directly, it took less than 0.2 second. I think how kylin try to
accomplish SUM(X) in a cube containing both TOPN(SUM(X)) and SUM(X) may not be
correct. That is the main reason I am against this decision that in a cube
containing TOPN(SUM(X)), one must also configure SUM(X)
#
> TopN requires a SUM to work
> ---------------------------
>
> Key: KYLIN-3322
> URL: https://issues.apache.org/jira/browse/KYLIN-3322
> Project: Kylin
> Issue Type: Bug
> Components: Measure - TopN
> Reporter: liyang
> Assignee: Na Zhai
> Priority: Major
>
> Currently if user creates a measure of TopN seller by sum of price, it is
> required that user also creates a measure of SUM(price). Otherwise, NPE will
> be thrown at query time.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
