[ 
https://issues.apache.org/jira/browse/HAWQ-1513?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

lynn updated HAWQ-1513:
-----------------------
    Description: 
SELECT
        mid,
        COUNT (mid),
        string_agg (
                create_time ,|| '#' || s_id
                ORDER BY
                        create_time
        )
FROM
        t1
WHERE
        t1.create_time BETWEEN to_timestamp(
                '2016-12-19 00:20:00:770',
                'yyyy-MM-dd HH24:MI:ss.ff'
        )
AND to_timestamp(
        '2016-12-19 23:40:00:770',
        'yyyy-MM-dd HH24:MI:ss.ff'
)group by mid;

when we explain the sql statement, we find it scan all the partition of table 
t1 in the query plan, and it has a huge impact on the query performance.
what can i do to solve this problem?
Thanks!

  was:
SELECT
        mid,
        COUNT (mid),
        string_agg (
                create_time ,|| '#' || s_id
                ORDER BY
                        create_time
        )
FROM
        t1
WHERE
        t1.create_time BETWEEN to_timestamp(
                '2016-12-19 00:20:00:770',
                'yyyy-MM-dd HH24:MI:ss.ff'
        )
AND to_timestamp(
        '2016-12-19 23:40:00:770',
        'yyyy-MM-dd HH24:MI:ss.ff'
)group by mid;

when we explain the sql statement, we find it scan all the partition of table 
t1 in the query plan, and it has a huge impact on the query performance.
what can i do to solve this problem?


> "string_agg" function does not support query optimization on  partitioned 
> table
> -------------------------------------------------------------------------------
>
>                 Key: HAWQ-1513
>                 URL: https://issues.apache.org/jira/browse/HAWQ-1513
>             Project: Apache HAWQ
>          Issue Type: Test
>          Components: Catalog
>            Reporter: lynn
>            Assignee: Radar Lei
>
> SELECT
>       mid,
>       COUNT (mid),
>       string_agg (
>               create_time ,|| '#' || s_id
>               ORDER BY
>                       create_time
>       )
> FROM
>       t1
> WHERE
>       t1.create_time BETWEEN to_timestamp(
>               '2016-12-19 00:20:00:770',
>               'yyyy-MM-dd HH24:MI:ss.ff'
>       )
> AND to_timestamp(
>       '2016-12-19 23:40:00:770',
>       'yyyy-MM-dd HH24:MI:ss.ff'
> )group by mid;
> when we explain the sql statement, we find it scan all the partition of table 
> t1 in the query plan, and it has a huge impact on the query performance.
> what can i do to solve this problem?
> Thanks!



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

Reply via email to