note 95033 deleted from function.ctype-digit by hholzgra

Note Submitter: walter ---- I was looking at whether this would save time on the numerous input validations I make in legions of scripts. Typically I use a function numbers_only() which simply does a preg_replace() to remove non-digits from a string. To test for a possible speedup, I created a new function which only performed the preg_replace() once a type_digit() check had failed. The results for 1 million interations showed that using ctype_digit() beforehand caused approximately 1/3rd additional latency on strings that were going to be preg_replace()'d anyway (ie: strings that did not contain only digits). It caused an over 100% speedup over that (latency inclusive) time for input strings that were pure numbers. The speedup was around 2/3 of the original blind preg_replace(). The lesson for me is that it's only worth trying to optimise away preg_replace() using ctype_digit() or similar if you know with some certainty that the vast majority of your inputs will lean one way or the other. Having said that, ctype_digit() seems to be cosnistently 30% faster than preg_match(). But adding the additional PHP option as a requirement for your codebase may not justify the optimisation. Let's face it: PHP aint exactly assembler, even if it is much faster than ruby :)