Comments on: PHP function that strips all non alphanumeric characters

By: Garrett

Garrett — Wed, 09 Dec 2009 15:14:27 +0000

Any idea which is faster?

preg_replace(’#[^a-z0-9]#i’, ”, $string)

preg_replace(’#[^a-z0-9A-Z]#’, ”, $string)

or do they compile down to the same regex and are equivalent?

Any slight improvement would make a huge difference in my script, as I’m running around 100 million replaces total.

By: Carsten

Carsten — Wed, 07 Oct 2009 19:54:45 +0000

\w and \W depend on the locale being used. This means that, for instance, with a German locale setting the code above will not remove ä ö ü Ä Ö Ü ß
Apart from that, \w always includes the underscore, which is not an alphanumeric.

preg_replace(’#[^a-z0-9]#i’, ”, $string) will do the job.