[ https://issues.apache.org/jira/browse/SPARK-10513?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14961740#comment-14961740 ]
Joseph K. Bradley commented on SPARK-10513: ------------------------------------------- Oh, I see, good question. I guess the root problem is that we should probably not allow empty feature names. How about we change StringIndexer to rename the empty string to "0" or the next available integer (or something like that)? We can document that behavior + note the behavior change in the release notes. * Later, we can add a Param to allow other behaviors: filter out Rows with invalid values, or throw an Exception. If this sounds reasonable, I can make a JIRA for it. > Springleaf Marketing Response > ----------------------------- > > Key: SPARK-10513 > URL: https://issues.apache.org/jira/browse/SPARK-10513 > Project: Spark > Issue Type: Sub-task > Components: ML > Reporter: Yanbo Liang > Assignee: Yanbo Liang > > Apply ML pipeline API to Springleaf Marketing Response > (https://www.kaggle.com/c/springleaf-marketing-response) -- 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