You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@kylin.apache.org by "Shaofeng SHI (JIRA)" <ji...@apache.org> on 2015/11/30 07:14:10 UTC

[jira] [Updated] (KYLIN-980) FactDistinctColumnsJob to support high cardinality columns

     [ https://issues.apache.org/jira/browse/KYLIN-980?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Shaofeng SHI updated KYLIN-980:
-------------------------------
    Fix Version/s: 2.0
      Component/s: Job Engine

Change the Mapper/combiner/reducer to leverage MapReduce to do the sort and group, then it doesn't need much physical memory.

The change is made on 2.x branch;

> FactDistinctColumnsJob to support high cardinality columns
> ----------------------------------------------------------
>
>                 Key: KYLIN-980
>                 URL: https://issues.apache.org/jira/browse/KYLIN-980
>             Project: Kylin
>          Issue Type: Improvement
>          Components: Job Engine
>    Affects Versions: v0.7.2
>            Reporter: Shaofeng SHI
>            Assignee: Shaofeng SHI
>              Labels: newbie
>             Fix For: 2.0
>
>
> In FactDistinctColumnsJob's combiner and reducer, it uses a HashSet to remove the duplicated values; But if a column's cardinality is very big, say > 10 Million, it may reports OutOfMemory error;
> It should be enhanced to support such case.



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