[ https://issues.apache.org/jira/browse/SPARK-4936?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14256985#comment-14256985 ]
Sean Owen commented on SPARK-4936: ---------------------------------- Are you referring to the NamedVector idea from Mahout? I think that already exists in a different form here. If you have an RDD of (<identifier>,Vector), then you can already use a clustering model to map all the values to a predicted cluster with mapValues(), and end up with an RDD of (<identifier>, <cluster>). If that's what you're looking for, then it does not require further work in Spark. > Please support Named Vector so as to maintain the record ID in clustering etc. > ------------------------------------------------------------------------------ > > Key: SPARK-4936 > URL: https://issues.apache.org/jira/browse/SPARK-4936 > Project: Spark > Issue Type: Improvement > Affects Versions: 1.1.1 > Reporter: mahesh bhole > Priority: Minor > > Hi > Please support Named Vector so as to maintain the record ID in clustering etc. > Thanks, > Mahesh -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org