> The most common problems are when processing CSVs produced elsewhere [...] The limitations of CSV are certainly worth considering and, in the instances you m

> The most common problems are when processing CSVs produced elsewhere [...] The limitations of CSV are certainly worth considering and, in the instances you mentioned, it may be not be worth using CSV. (If you are going to be using a more complex parser anyway, you may as well using a format that is better defined and where you are less likely to encounter edge cases.) That being said, there remain many cases where CSV is far more efficient and far less error prone.