We've since fixed the issue with the CSV parsing so it should work correctly with the example provided in the documentation. Thanks for the report
