preg_match_all - exclude a word from the search php

Question

I want to exclude a class from this regex, for example "apple" or "apple-tree"

preg_match_all('/class[~=]+[\'"](.*?)[\'"]/i', $html, $matches1, PREG_SET_ORDER, 0);

How would I achieve this in php regex?

Answer 1

You can do this with a negative lookahead, BUT it's a slippery slope and I recommend using something other than regular expressions.

Let's start here (removing the unrelated regex for now, just for brevity):

class="((?!apple).*?)"

This looks good at first glance; it successfully ignores classes that equal or start with apple. But what if we want to ignore matches that have apple anywhere in the list? Because the class attribute can have many classes in any particular order.

class="((?:(?!apple).)*?)"

Nice, that takes care of excluding a class attribute where apple appears anywhere in the value. (More in-depth explanation of this.)

But, crap, our important customer Snapple reported a bug where their custom class names aren't working anymore.

You can see where this is going.

It's very possible that you have a more controlled scenario and one of the above will be useful. If so, awesome! I hope this helps.

For some alternatives to trying to do this by regular expressions, consider:

• Loading and traversing the HTML via DOMDocument
• Similarly, using DOMDocument + Xpath querying (you can do some really fun things with Xpath)
• For something more simple that doesn't seem to warrant all that nonsense, something like preg_replace_callback(), where you can add in some business logic around a more generic matching pattern