Every data processing task has to make assumptions about the well-formedness of its input. "Arbitrary CSV" is basically undefined; whether deviations are best d

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.