org.apache.spark.sql.connector.V2CommandsCaseSensitivitySuite.ReplaceTableAsSelect: using top level field for partitioning