How well do those tools work with arbitrary CSV files, e.g. containing line breaks or quotes in field data? I wasn't aware that they can actually parse CSV and instead you have to assume things about the content that may not end up being true.
Every data processing task has to make assumptions about the well-formedness of its input. "Arbitrary CSV" is basically undefined; whether deviations are best dealt with by parsing, preprocessing, or different tools altogether depends on the source.
