[jira] [Created] (KYLIN-2948) Count a column returns the same result as count(*) even if this column has NULL

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

DeXin created KYLIN-2948:
----------------------------

             Summary: Count a column returns the same result as count(*) even if this column has NULL
                 Key: KYLIN-2948
                 URL: https://issues.apache.org/jira/browse/KYLIN-2948
             Project: Kylin
          Issue Type: Bug
    Affects Versions: v2.1.0
         Environment: CentOS 7
            Reporter: DeXin
            Priority: Critical


When we want to count a column(with same NULL value), there is different result from kylin and hive SQL. Is there a way to exclude NULL value in count measure calculation for a particular column? 
Here is the example:

1. Here is source data:
Date             ID
2017-10-10  dfe343ddfe3f5
2017-10-11  fer234d656dff
2017-10-11  NULL
2017-10-12  jui6jnc3ncce3

2. run SQL in Hive:
select Date, count(*), count(ID) from table group by Date;
2017-10-10  1  1
2017-10-11  2  1
2017-10-12  1  1

3. run same SQL in Kylin:
select Date, count(*), count(ID) from table group by Date;
2017-10-10  1  1
2017-10-11  2  2
2017-10-12  1  1



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