[jira] [Updated] (HIVE-4109) Partition by column does not have to be in order by

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

     [ https://issues.apache.org/jira/browse/HIVE-4109?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Brock Noland updated HIVE-4109:
-------------------------------

    Description: 
Cam up in the review of HIVE-4093.

Ashutosh
{noformat}
I am not sure if this is illegal query. I tried following two queries in postgres, both of them succeeded.

select p_mfgr, avg(p_retailprice) over(partition by p_mfgr, p_type order by p_mfgr) from part;
select p_mfgr, avg(p_retailprice) over(partition by p_mfgr order by p_type,p_mfgr) from part;
{noformat}

Harish
{noformat}
The first one doesn't make sense, right? Order on a subset of the partition columns

The second one: Can we do this with the Hive ReduceOp have the orderColumns be in a different order than the key columns?
{noformat}

  was:
Cam up in the review of HIVE-4093.

{noformat}
I am not sure if this is illegal query. I tried following two queries in postgres, both of them succeeded.

select p_mfgr, avg(p_retailprice) over(partition by p_mfgr, p_type order by p_mfgr) from part;
select p_mfgr, avg(p_retailprice) over(partition by p_mfgr order by p_type,p_mfgr) from part;
{noformat}

    
> Partition by column does not have to be in order by
> ---------------------------------------------------
>
>                 Key: HIVE-4109
>                 URL: https://issues.apache.org/jira/browse/HIVE-4109
>             Project: Hive
>          Issue Type: Bug
>          Components: PTF-Windowing
>            Reporter: Brock Noland
>
> Cam up in the review of HIVE-4093.
> Ashutosh
> {noformat}
> I am not sure if this is illegal query. I tried following two queries in postgres, both of them succeeded.
> select p_mfgr, avg(p_retailprice) over(partition by p_mfgr, p_type order by p_mfgr) from part;
> select p_mfgr, avg(p_retailprice) over(partition by p_mfgr order by p_type,p_mfgr) from part;
> {noformat}
> Harish
> {noformat}
> The first one doesn't make sense, right? Order on a subset of the partition columns
> The second one: Can we do this with the Hive ReduceOp have the orderColumns be in a different order than the key columns?
> {noformat}

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira