Re: Breaking up data efficiently

> In my opinion you should save exceptions for, well, EXCEPTIONal > problems. Either that or you use exceptions for a large system that > discriminates against specific types of errors and handles each error > type in a totally different way. Can we say bloat? Depending on the system, malformatted data may very well be exceptional. (adj. 1. being an exception; uncommon) Judging by the lack of error-checking in the OP, it seems to be the case. Using trigger_error or return false would both be reasonable solutions as well. Really, how errors are handled is up to the individual. My point was that Wee Keat may consider adding a few lines for validation. Certainly adding any error-checking will reduce efficiency, which is what the original question was regarding. (As an aside, some day i will have to find out how many times the word bloat appears on this list.)