RE: [PHP] Filters and sanitizing a regexp

> -----Original Message----- > From: richard.heyes@gmail.com [mailto:richard.heyes@gmail.com] On > Behalf Of Richard Heyes > Sent: Thursday, September 25, 2008 2:28 PM > To: Boyd, Todd M. > Cc: Frank Stanovcak; php-general@lists.php.net > Subject: Re: [PHP] Filters and sanitizing a regexp > > > ... > > Hi, > > Had to have a play with this. Here's my (somewhat stricter) regex: > > preg_match('/((?:[0-3]?[0-9])?\d)-([0-1]?\d)-20\d\d/', '31-12-2000', > $matches); > if (!empty($matches[1]) AND $matches[1] > 31) $matches = array(); > if (!empty($matches[2]) AND $matches[2] > 12) $matches = array(); > > You could probably do the checks for the day and month within the > regex, but it's late :-) I don't think that one is quite right... Your optionals (?) are confusing. Why [0-3]?[0-9] AND the \d? Why [0-9] at all (since \d is essentially [0-9])? Why use the discarded capture group (?:) since the [0-3] is already optional? It's on! :D <?php $var = "as12/32/2000df"; $matches = array(); preg_match('#(\d?[0-9])\D*([0-3]?\d)\D*(\d{4})#', $var, $matches); if(count($matches) != 4 || $matches[1] > 12 || $matches[2] > 31) $matches = NULL; ?> MMDDYYYY 4 life, yo! Although, I would honestly prefer it if the world went to YYYYMMDD as a standard. Just seems to sort itself out more intuitively to the way I perceive time. (FYI, I'm not sure about the if() statement working out or not... I'm assuming PHP will drop through to the "then" part if there aren't 4 elements in the array instead of arbitrarily testing the possibly-missing elements... but I didn't test that part.) Also, I used \D* to save space (and confusion). You could just as easily use (?:-|/) for hyphens or slashes. Todd Boyd Web Programmer