unsubscribe
Hi,
We have a usecase where we are submitting multiple spark jobs using
SparkLauncher from a Java class.
We are currently in a memory crunch situation on our edge node where we see
that the Java processes spawned by the launcher is taking around 1 GB.
Is there a way to pass JMX parameters to this
Hi all,
I'm trying to group X rows in a single one without shuffling the date.
I was thinking doing something like that :
val myDF = Seq(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11).toDF("myColumn")
myDF.withColumn("myColumn", expr("sliding(myColumn, 3)"))
expected result:
myColumn
[1,2,3]
[4,5,6]
Greetings,
tl;dr there must have been a regression in spark *connect*'s ability to
retrieve data, more details in linked issues
https://issues.apache.org/jira/browse/SPARK-45598
https://issues.apache.org/jira/browse/SPARK-45769
we have projects that depend on spark connect 3.5 and we'd