[ https://issues.apache.org/jira/browse/SPARK-5987?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14357109#comment-14357109 ]
Joseph K. Bradley commented on SPARK-5987: ------------------------------------------ This isn't a bug in Spark SQL. The issue is that we haven't defined a UserDefinedType for Matrices. (We should, but haven't yet.) When I said "basic types," I meant the types enumerated on the SQL programming guide (basically, Array[Double] or Seq[Double] will be best). I'd recommend flattening the matrix into an Array[Double] instead of having nested types. The nesting is less efficient because of all of the extra objects it creates. > Model import/export for GaussianMixtureModel > -------------------------------------------- > > Key: SPARK-5987 > URL: https://issues.apache.org/jira/browse/SPARK-5987 > Project: Spark > Issue Type: Sub-task > Components: MLlib > Affects Versions: 1.3.0 > Reporter: Joseph K. Bradley > Assignee: Manoj Kumar > > Support save/load for GaussianMixtureModel -- 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