[ https://issues.apache.org/jira/browse/SPARK-12577?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15083590#comment-15083590 ]
Thomas Sebastian edited comment on SPARK-12577 at 1/5/16 7:41 PM: ------------------------------------------------------------------ Hi Reynold, Would you share some thoughts on how did you replicate this issue? - which version of spark.? - A bit more failure message details( what sort of exception).? Do you mean to say, when the sqlContext based queries(spark-shell) are fired as in the above FAIL conditions, it does not go through, where as it is accepted via HiveQL ? Also, I see some paranthesis close mismatch in the PASS conditions mentioned. was (Author: thomastechs): Hi Reynold, Would you share some thoughts on how did you replicate this issue? - using sqlContext or API? - which version of spark.? - A bit more failure message details( what sort of exception).? Also, I see some paranthesis close mismatch in the PASS conditions mentioned. > better support of parentheses in partition by and order by clause of window > function's over clause > -------------------------------------------------------------------------------------------------- > > Key: SPARK-12577 > URL: https://issues.apache.org/jira/browse/SPARK-12577 > Project: Spark > Issue Type: Sub-task > Components: SQL > Reporter: Reynold Xin > > Right now, Hive's parser support > {code} > -- PASS > SELECT SUM(1) OVER (PARTITION BY a + 1 - b * c / d FROM src; > SELECT SUM(1) OVER (PARTITION BY (a + 1 - b * c / d) FROM src; > {code} > But, the following one is not accepted > {code} > -- FAIL > SELECT SUM(1) OVER (PARTITION BY (a) + 1 - b * c / d) FROM src; > {code} > We should fix it in our own parser. -- 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