[
https://issues.apache.org/jira/browse/MAHOUT-352?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12851468#action_12851468
]
Sean Owen commented on MAHOUT-352:
----------------------------------
The output compresses well, so I think it's important to save the I/O to HDFS
by compressing. I suppose a caller might frequently only request
recommendations for a limited number of users, so it's not much of an issue. I
suppose it's not hard to make this an option, though the number of arguments is
starting to get large.
> not compression the final output of RecommenderJob
> --------------------------------------------------
>
> Key: MAHOUT-352
> URL: https://issues.apache.org/jira/browse/MAHOUT-352
> Project: Mahout
> Issue Type: Wish
> Components: Collaborative Filtering
> Affects Versions: 0.4
> Reporter: Hui Wen Han
> Priority: Minor
>
> in RecommenderJob Class:
> line 113 :recommenderConf.setClass("mapred.output.compression.codec",
> GzipCodec.class, CompressionCodec.class);
> because the output here is the final output ,the end user may operates the
> output through HDFS API,
> if compress here ,it not very convenient.
> but it's convenient for operation through HDFS　shell to get to local.
> can provider options or not compress ?
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
