Unable to do count or isempty call in dataframe with exceptall

Hey,

Basically this is a bug in spark 3.3.0 as mentioned in this jira - https://issues.apache.org/jira/browse/SPARK-39612 caused by Prune unrequired child index - https://issues.apache.org/jira/browse/SPARK-38531.

I tested and verified at my end as well. The fix is available in park 3.3.1, 3.4.0. EMR 6.10.1 release has spark 3.3.1 that natively support this. - https://docs.aws.amazon.com/emr/latest/ReleaseGuide/emr-6101-release.html