This is not making it consistent when we perform aggregation on multiple columns and thus it is hindering metadata driven transformation because of inconsistency.
How can we request Databricks/pyspark to include this ? and is there any known work around meanwhile ?
