multiple column distinct count

[市场中心-ZHANGDA32698 <zd...@ly.com>]

Hi,

As stated in the release note and jira https://issues.apache.org/jira/browse/KYLIN-490 , multiple column distinct count is supported in v2.0. So in order to do ‘select count distinct (A,B) from table ’, I assume I need to specify a count-distinct measure that includes both A and B in the cube design stage, but I notice that in the ‘edit measure’ UI, the ‘Param Value’ is a drop-down list where I can’t input more than 1 column. I’m curious how does kylin do multi column count distinct query without defining a multi-column count-distinct measure?

Re: multiple column distinct count

[Li Yang <li...@apache.org>]

You are right. The GUI still cannot input multiple columns for the
count-distinct measure. A JIRA is created.
https://issues.apache.org/jira/browse/KYLIN-2616



On Thu, May 4, 2017 at 6:04 PM, 市场中心-ZHANGDA32698 <zd...@ly.com> wrote:

> Hi,
>
>
>
> As stated in the release note and jira https://issues.apache.org/
> jira/browse/KYLIN-490 , multiple column distinct count is supported in
> v2.0. So in order to do ‘select count distinct (A,B) from table ’, I assume
> I need to specify a count-distinct measure that includes both A and B in
> the cube design stage, but I notice that in the ‘edit measure’ UI, the
> ‘Param Value’ is a drop-down list where I can’t input more than 1 column.
> I’m curious how does kylin do multi column count distinct query without
> defining a multi-column count-distinct measure?
>