Aniket Arun Kulkarni created SPARK-23876:
--------------------------------------------
Summary: OR condition in joins causes results to come back to driver
Key: SPARK-23876
URL: https://issues.apache.org/jira/browse/SPARK-23876
Project: Spark
Issue Type: Bug
Components: Spark Core
Affects Versions: 1.6.2
Reporter: Aniket Arun Kulkarni
Hello all,
I am trying to implement an 'OR' join, something as shown below,
leftDF.join(rightDF, leftDF(leftJoinKey1) === rightDF(rightJoinKey1) || leftDF(leftJoinKey2)
=== rightDF(rightJoinKey2), joinType)
.select(leftDF.col("*"), rightDF(colToFetchFromRight) as(aliasForRightColumn))
rightDF has around 400 million rows
leftDF has around 200 million rows
When I run this code, I get an error saying
"Total size of serialized results is bigger than spark.driver.maxResultSize"
This means the results are being sent back to the driver even when i did not call actions
such as count or show.
If I remove the or condition, the results are no longer sent to the driver and the job goes
through.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org
|