[
https://issues.apache.org/jira/browse/SPARK-29265?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16939381#comment-16939381
]
Florentino Sainz commented on SPARK-29265:
------------------------------------------
Ok, after quite a bit of research, I had to change my Window to groupBy +
collect_list
Unbounded+unordered Window + collect_list seems to be collecting ALL the
dataframe in a single executor/task.
groupBy + collect_list seems to do it as expect (collect_list for each group
independently).
Is this behaviour expected or it's a bug?
> Window orderBy causing full-DF orderBy
> ---------------------------------------
>
> Key: SPARK-29265
> URL: https://issues.apache.org/jira/browse/SPARK-29265
> Project: Spark
> Issue Type: Bug
> Components: Spark Core
> Affects Versions: 2.4.0
> Environment: Any
> Reporter: Florentino Sainz
> Priority: Minor
>
> Hi,
>
> I had this problem in "real" environments and also made a self-contained test
> ( [^Test.scala] attached).
> Having this Window definition:
> {code:scala}
> val myWindow = Window.partitionBy($"word").orderBy("word")
> val filt2 = filtrador.withColumn("avg_Time",
> collect_list($"number").over(myWindow))
> {code}
>
> As a user, I would expect either:
> 1- Error/warning (because trying to sort on one of the columns of the window
> partitionBy)
> 2- A mostly-useless operation which just orders the rows (which have the same
> value on word) inside each Window but doesn't affect performance too much.
>
> Currently what I see:
> *When I use "myWindow" in any DataFrame, somehow that Window.orderBy is
> performing a global orderBy on the whole DataFrame. Similar to
> dataframe.orderBy("word").*
> *In my real environment, my program just didn't finish in time/crashed thus
> causing my program to be very slow or crash (because as it's a global
> orderBy, it will just go to one executor).*
>
> In the test I can see how all elements of my DF are in a single partition
> (side-effect of the global orderBy)
>
> Full Code showing the error (see how the mapPartitions shows 99 rows in one
> partition) attached in Test.scala, sbt project (src and build.sbt) attached
> too in TestSpark.zip.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
