[jira] [Commented] (KYLIN-3463) Improve optimize job by avoiding creating empty output files on HDFS

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

    [ https://issues.apache.org/jira/browse/KYLIN-3463?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16552397#comment-16552397 ] 

Zhong Yanghong commented on KYLIN-3463:
---------------------------------------

Hi [~kangkaisen], could you help review this patch? It's similar to {{FactDistinctColumnsJob}}. Thanks very much.

> Improve optimize job by avoiding creating empty output files on HDFS
> --------------------------------------------------------------------
>
>                 Key: KYLIN-3463
>                 URL: https://issues.apache.org/jira/browse/KYLIN-3463
>             Project: Kylin
>          Issue Type: Improvement
>            Reporter: Zhong Yanghong
>            Assignee: Zhong Yanghong
>            Priority: Major
>         Attachments: APACHE-KYLIN-3463.patch
>
>
> For steps in an optimize job, like {{FilterRecommendCuboidDataJob}} & {{UpdateOldCuboidShardJob}}, {{MultipleOutputs}} is used to output data into different directories. Therefore, the default output file is empty. {{LazyOutputFormat}} should be used to prevent to create zero-sized default output.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)