[
https://issues.apache.org/jira/browse/SPARK-11027?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Apache Spark reassigned SPARK-11027:
------------------------------------
Assignee: (was: Apache Spark)
> Better group distinct columns in query compilation
> --------------------------------------------------
>
> Key: SPARK-11027
> URL: https://issues.apache.org/jira/browse/SPARK-11027
> Project: Spark
> Issue Type: Sub-task
> Components: SQL
> Reporter: Yin Huai
>
> In AggregationQuerySuite, we have a test
> {code}
> checkAnswer(
> sqlContext.sql(
> """
> |SELECT sum(distinct value1), kEY - 100, count(distinct value1)
> |FROM agg2
> |GROUP BY Key - 100
> """.stripMargin),
> Row(40, -99, 2) :: Row(0, -98, 2) :: Row(null, -97, 0) :: Row(30, null,
> 3) :: Nil)
> {code}
> We will treat it as having two distinct columns because sum causes a cast on
> value1. Maybe we can ignore the cast when we group distinct columns. So, it
> will not be treated as having two distinct columns.
--
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
