Question

I’m having a bit of difficulties converting some regex from being used in preg_match_all to being used in preg_replace.

Basically, via regex only, I would like to match uppercase characters that are preceded by either a space, beginning of text, or a hypen. This is not a problem, I have the following for this which works well:

preg_match_all('/(?<= |\A|-)[A-Z]/',$str,$results);
echo '<pre>' . print_r($results,true) . '</pre>';

Now, what I’d like to do, is to use preg_replace to only return the string with the uppercase characters that match my criteria above. If I port the regex straight into preg_replace, then it obviously replaces the characters I want to keep.

Any help would be much appreciated 🙂

Also, I’m fully aware regex isn’t the best solution for this in terms of efficiency, but nonetheless I would like to use preg_replace.

Answer 1

According to De Morgan’s laws,
if you want to keep letters that are

• A-Z, and
• preceded by [space], \A, or -

then you’d want to remove characters that are

• not A-Z, or
• not preceded by [space], \A, or -

Perhaps this (replace match with empty string)?

/[^A-Z]|(?<! |\A|-)./

See example here.