When you run a mapping on the Spark engine, you can read data from and write data to Avro, ORC, and Parquet files that are partitioned based on directories.

You can read and write flat files with and without headers.

You can configure row delimiter and define qualifier scope when you read and write flat files.