[ https://issues.apache.org/jira/browse/SPARK-4844?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14246738#comment-14246738 ]
Guoqiang Li commented on SPARK-4844: ------------------------------------ The main reason is that {{RDD.sample}} is not efficient. {{RDD.sample}} loads all data into memory. See https://github.com/witgo/spark/compare/SPARK-4844 > SGD should support custom sampling. > ----------------------------------- > > Key: SPARK-4844 > URL: https://issues.apache.org/jira/browse/SPARK-4844 > Project: Spark > Issue Type: Improvement > Components: MLlib > Reporter: Guoqiang Li > Fix For: 1.3.0 > > -- 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