org.apache.spark.sql.execution
Find out duplicated exchanges in the spark plan, then use the same exchange for all the references.
Find out duplicated exchanges in the spark plan, then use the same exchange for all the references.