Remembering why I hate Regex

RobertK · 01-08-2008, 09:33 PM

Well, I'm trying to do a quick regex match here, and it won't stop choking. The strings are formatted as:

abc
abc123
abc123,1

I am trying to skip the initial text, if any, and find that middle number before the comma. Unfortunately it keeps dying. The center number can reach from 1-3 digits in length.

I've tried patterns based off this:

PHP Code:


			
  preg_match('/([\d]{1,3}*?)[,]*/', $hash, $matches);

  if(count($matches) === 1) {

    if(is_numeric($matches[0][0])) {

      $bits = intval($matches[0][0]);

      $hex = $bits / 4;

    }

  }

  unset($matches);

However it tells me: "Compilation failed: nothing to repeat at offset 10"

I hope someone around here can remind me how to do it right...

Wildhoney · 01-08-2008, 09:41 PM

Would the following be OK?

php Code:

$szText = 'abc123,1';
preg_match('~(?P<numbers>[0-9]{1,3})~', $szText, $aMatches);

var_dump((int) $aMatches['numbers']);

RobertK · 01-08-2008, 09:52 PM

Yeah, thanks. That helps a lot. Found around 2-3 different errors in that section total. Now I'm just working on a switch statement that isn't working.

Salathe · 01-09-2008, 02:26 PM

You've already got the problem solved but I see no reason to waste time and energy when it's not necessary. Is there any reason why the following would not suffice?

PHP Code:


			
$hash = 'abc123,1';
if (preg_match('/\d{1,3}/', $hash, $matches))
{
    $bits = (int) $matches[0];
    $hex  = $bits / 4;
    unset ($matches);
}

RobertK · 01-09-2008, 02:37 PM

Methinks that would work a bit faster, without the string match. Thanks Salathe.