You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hive.apache.org by "Xuefu Zhang (JIRA)" <ji...@apache.org> on 2014/07/29 15:13:39 UTC

[jira] [Commented] (HIVE-7328) Refactoring: make Hive reduce side data processing reusable

    [ https://issues.apache.org/jira/browse/HIVE-7328?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14077694#comment-14077694 ] 

Xuefu Zhang commented on HIVE-7328:
-----------------------------------

It seems it's easier to use ExecReducer directly than any refactoring. Postpone this item for now for later consideration.

> Refactoring: make Hive reduce side data processing reusable
> -----------------------------------------------------------
>
>                 Key: HIVE-7328
>                 URL: https://issues.apache.org/jira/browse/HIVE-7328
>             Project: Hive
>          Issue Type: Sub-task
>    Affects Versions: 0.13.0
>            Reporter: Xuefu Zhang
>            Assignee: Xuefu Zhang
>
> ExecReducer is Hive's reducer implementation for MapReduce. Table rows are shuffled by MR framework to ExecReducer and further processed by ExecReducer.reduce() method, which invokes Hive's reduce-side operator tree starting. This task is to extract the reduce-side data processing offered by the operator tree so that it can be reused by other execution engine such as Spark. This is purely refactoring the existing code.



--
This message was sent by Atlassian JIRA
(v6.2#6252)