Posted by Rik on 05/10/06 06:35

Jerry Fleming wrote:
> Hi,
>
> I want to strip part of a string that *doesn't* match a patter. But
> preg_replace does the opposite. For example:
>
> $str = 'aa bb';
> echo preg_replace('/bb|cc/', '', $str);
>
> where aa is unpredictable, so my pattern can only be bb|cc. The
> problem is, how can I replace aa by negating a pattern?


Very difficult with preg_replace. If you know _exactly_ what is allowed, why
not preg_match() the string and continue working with the output from that
function? Negating characters, getting parts NOT followed or predeeded by a
specific string can be done, but negating whole words is as far as I know
impossible.

For instance:

Allowed are 'cat' & 'dog' & whitespacecharacters.
$tring = "dog dogcat\ncowcat";
preg_match_all('/(cat|dog|\s)/si', $tring, $matches, PREG_PATTERN_ORDER);
$allowedstring = implode($matches[1]);

Grtz,
--
Rik Wasmus

[Back to original message]