[
https://issues.apache.org/jira/browse/KYLIN-5355?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17647067#comment-17647067
]
pengfei.zhan edited comment on KYLIN-5355 at 12/14/22 1:32 PM:
---------------------------------------------------------------
{code}
select max(1) from table
{code}
When matching models, this kind of statement will get an empty realization of
IRealization, the spark plan throws NPE if the realization is empty.
was (Author: JIRAUSER294653):
How to reproduce this problem?
1) create a model, and add some dimensions(the more the better) and measures,
2) add two aggregation groups, one group is a sub-group of another, i.e. the
dimensions of the first group should contain all the dimensions of the second,
note that the dimensions order is not the same. For example, the group1 has
dimensions with an order of [a, b, c, d, e, m1, m2], and the group2 has
dimensions with an order of [b, e, a, m1, m2].
3) save the changes.
> Support constant aggregate on a table
> -------------------------------------
>
> Key: KYLIN-5355
> URL: https://issues.apache.org/jira/browse/KYLIN-5355
> Project: Kylin
> Issue Type: Improvement
> Components: Query Engine
> Affects Versions: 5.0-alpha
> Reporter: pengfei.zhan
> Assignee: pengfei.zhan
> Priority: Major
> Fix For: 5.0-alpha
>
>
> At present, KYLIN supports some constant aggregate as below.
> {code:sql}
> select sum(1), min(1), max(1);
> {code}
> However, this kind of constant aggregate as below doesn't support query by
> model.
> {code:sql}
> select sum(1), min(1), max(1) from table;
> {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
