In general, yes. But, given that different applications and people may have varying levels of strictness about their CSV-ish-ness, I would do some testing before I commit to automating in production or something like that. Even after that, you should try to think of some (all?) of the ways that the creator of the CSV file can screw up the format and try to detect them.