[ https://issues.apache.org/jira/browse/SPARK-11288?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15011855#comment-15011855 ]
Jakob Odersky commented on SPARK-11288: --------------------------------------- Tell me if I'm missing the point, but you can also pass type parameters explicitly when calling udfs: {code} df.udf[ReturnType, Arg1, Arg2]((arg1, arg2) => ret) {code} > Specify the return type for UDF in Scala > ---------------------------------------- > > Key: SPARK-11288 > URL: https://issues.apache.org/jira/browse/SPARK-11288 > Project: Spark > Issue Type: New Feature > Components: SQL > Reporter: Davies Liu > > The return type is figured out from the function signature, maybe it's not > that user want, for example, the default DecimalType is (38, 18), user may > want (38, 0). > The older deprecated one callUDF can do that, we should figure out a way to > support that. > cc [~marmbrus] -- 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