[jira] [Updated] (MAHOUT-1637) RecommenderJob of ALS fails in the mapper because it uses the instance of other class

["Cristian Galán (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/MAHOUT-1637?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Cristian Galán updated MAHOUT-1637:
-----------------------------------
    Summary: RecommenderJob of ALS fails in the mapper because it uses the instance of other class  (was: RecommenderJob of ALS fails in the mapper because uses the instance of other class)

> RecommenderJob of ALS fails in the mapper because it uses the instance of other class
> -------------------------------------------------------------------------------------
>
>                 Key: MAHOUT-1637
>                 URL: https://issues.apache.org/jira/browse/MAHOUT-1637
>             Project: Mahout
>          Issue Type: Bug
>          Components: Collaborative Filtering
>    Affects Versions: 0.9
>            Reporter: Cristian Galán
>              Labels: als, collaborative-filtering
>             Fix For: 1.0
>
>   Original Estimate: 4h
>  Remaining Estimate: 4h
>
> In the map method of PredictionMapper when executes the next line:       Pair<OpenIntObjectHashMap<Vector>, OpenIntObjectHashMap<Vector>> uAndM = getSharedInstance();
> the job fail because obtains the instance of other class. This occurs because I launch a local job, so the instance exists previously and for this doesn't make the new correct instance for ALS.
> The solution that it works me is to add the next line:.
>         SharingMapper.reset();
> in method run of JobRecommender of org.apache.mahout.cf.taste.hadoop.als package 
> I have to test it in my environment with distributed mapreduce, hadoop fs, zookeeper and others if it works correctly.



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