Extend PARSE with a CSV(...) option. The idea is to expose the CSV parse for dataset to a broader use.
Some CSV files have sequences that prevent the file from being sprayed correctly and parsed correctly. In particular, embedded line end sequences in text fields can pose difficulties. One means of fixing these files is to process the stream and re-assemble the broken strings. At that point it would be convenient to be able to use the dataset CSV parse feature.
There are 2 current work arounds available. The first is to just write the file with a OUTPUT(...CSV) action. The second option is the use PARSE and a specific rule to perform the field assignments.