The DCA reads and writes csv's. These need to be consistently processed to avoid downstream inconsistencies in csv standard. The read/write operations also need to be efficient since some of our csv's contain millions of records.