[
https://issues.apache.org/jira/browse/SPARK-16642?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15385569#comment-15385569
]
Apache Spark commented on SPARK-16642:
--------------------------------------
User 'yhuai' has created a pull request for this issue:
https://github.com/apache/spark/pull/14284
> ResolveWindowFrame should not be triggered on UnresolvedFunctions.
> ------------------------------------------------------------------
>
> Key: SPARK-16642
> URL: https://issues.apache.org/jira/browse/SPARK-16642
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Reporter: Yin Huai
> Assignee: Yin Huai
>
> The case at
> https://github.com/apache/spark/blob/75146be6ba5e9f559f5f15430310bb476ee0812c/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala#L1790-L1792
> is shown below
> {code}
> case we @ WindowExpression(e, s @ WindowSpecDefinition(_, o,
> UnspecifiedFrame)) =>
> val frame = SpecifiedWindowFrame.defaultWindowFrame(o.nonEmpty,
> acceptWindowFrame = true)
> we.copy(windowSpec = s.copy(frameSpecification = frame))
> {code}
> This case will be triggered even when the function is an unresolved. So, when
> the functions like lead are used, we may see errors like {{Window Frame RANGE
> BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW must match the required frame
> ROWS BETWEEN 1 FOLLOWING AND 1 FOLLOWING.}} because we wrongly set the the
> frame specification.
--
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
