hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ashutosh Chauhan (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HIVE-9618) Deduplicate RS keys for ptf/windowing
Date Mon, 09 Feb 2015 19:20:36 GMT

    [ https://issues.apache.org/jira/browse/HIVE-9618?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14312684#comment-14312684
] 

Ashutosh Chauhan commented on HIVE-9618:
----------------------------------------

{code}
+      int index = ExprNodeDescUtils.indexOf(colDef.getExprNode(), partCols);
{code}

should this be instead :
{code}
+      int index = ExprNodeDescUtils.indexOf(colDef.getExprNode(), orderCols);
{code}

since idea is not to add same columns twice in orderCols, right ?

> Deduplicate RS keys for ptf/windowing
> -------------------------------------
>
>                 Key: HIVE-9618
>                 URL: https://issues.apache.org/jira/browse/HIVE-9618
>             Project: Hive
>          Issue Type: Improvement
>          Components: PTF-Windowing
>            Reporter: Navis
>            Assignee: Navis
>            Priority: Trivial
>         Attachments: HIVE-9618.1.patch.txt
>
>
> Currently, partition spec containing same column for partition-by and order-by makes
duplicated key column for RS. For example, 
> {noformat}
> explain
> select p_mfgr, p_name, p_size, 
> rank() over (partition by p_mfgr order by p_name) as r, 
> dense_rank() over (partition by p_mfgr order by p_name) as dr, 
> sum(p_retailprice) over (partition by p_mfgr order by p_name rows between unbounded preceding
and current row)  as s1
> from noop(on noopwithmap(on noop(on part 
> partition by p_mfgr 
> order by p_mfgr, p_name
> )))
> {noformat}
> "partition by p_mfgr order by p_mfgr, p_name" makes duplicated key columns like below
> {noformat}
> Reduce Output Operator
>     key expressions: p_mfgr (type: string), p_mfgr (type: string), p_name (type: string)
>     sort order: +++
>     Map-reduce partition columns: p_mfgr (type: string)
>     value expressions: p_size (type: int), p_retailprice (type: double)
> {noformat}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message