[jira] [Commented] (GIRAPH-1161) implement random sampling for input splits

["ASF GitHub Bot (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/GIRAPH-1161?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16186140#comment-16186140 ] 

ASF GitHub Bot commented on GIRAPH-1161:
----------------------------------------

GitHub user JianlongZhong opened a pull request:

    https://github.com/apache/giraph/pull/50

    GIRAPH-1161 - implement random sampling for input splits

    

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/JianlongZhong/giraph random_sample

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/giraph/pull/50.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #50
    
----
commit 0c285d3c86d10ad6084e65f00e8c2d45fbeea8ba
Author: Jianlong Zhong <ji...@fb.com>
Date:   2017-09-29T17:36:27Z

    GIRAPH-1161 - implement random sampling for input splits

----


> implement random sampling for input splits
> ------------------------------------------
>
>                 Key: GIRAPH-1161
>                 URL: https://issues.apache.org/jira/browse/GIRAPH-1161
>             Project: Giraph
>          Issue Type: Improvement
>            Reporter: Jianlong Zhong
>            Priority: Minor
>
> Currently if we are reading vertex/edge data from multiple tables, and we only want to read a fraction of data (with giraph.inputSplitSamplePercent conf option), we'll always get the first inputSplitSamplePercent of the input slits. We should instead use a random sample of input splits so testing on sample of data would look closer to actual full data run.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)