That is a pretty big assumption. Big assumptions like that are what leads to fragile infrastructures. Defensive programmers try to avoid that kind of fragility if at all possible...

Well, in this case, that he chose those for his delimiters would lead me to believe that his data is what I might call 'vanilla' (no field contains a pipe or newline). Besides which, once he puts the data in to a RDBMS, thenhe doesn't have to worry about it.

Just out or curiosity, what do you consider good delimiters? I use '||' for field delimiters, and '@@\n" for row delimiters...

I consider going to fancier delimiters because your old ones broke a design mistake. The problem isn't that there is a right delimiter. The problem is that you are creating data formats with known limitations, and those limitations are going to bite you.

I rather prefer data formats with provisions for escaping mechanisms allowing them to handle arbitrary data. Nothing less will suffice to allow for reliable software.