[ https://issues.apache.org/jira/browse/SPARK-14666?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15243455#comment-15243455 ]
Dominic Ricard edited comment on SPARK-14666 at 4/15/16 7:33 PM: ----------------------------------------------------------------- It appears that the order by clause is expecting the column name. The proper query looks like this: {noformat} select distinct concat(column_1, ' : ', column_2) as `column1` from table order by `column1`; {noformat} Why isn't concat(column_1, ' : ', column2) converted to the _c0 alias in order by when used in conjunction with distinct? was (Author: dricard): It appears that the order by clause is expecting the column name. The proper query looks like this: {noformat} select distinct concat(column_1, ' : ', column_2) as `column1` from table order by `column1`; {noformat} > Using DISTINCT on a UDF (like CONCAT) is not supported > ------------------------------------------------------ > > Key: SPARK-14666 > URL: https://issues.apache.org/jira/browse/SPARK-14666 > Project: Spark > Issue Type: Bug > Components: SQL > Affects Versions: 1.6.1 > Reporter: Dominic Ricard > Priority: Minor > > The following query fails with: > {noformat} > Java::JavaSql::SQLException: org.apache.spark.sql.AnalysisException: cannot > resolve 'column_1' given input columns: [_c0]; line # pos ## > {noformat} > Query: > {noformat} > select > distinct concat(column_1, ' : ', column_2) > from > table > order by > concat(column_1, ' : ', column_2); > {noformat} -- 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