Regex Multiple Capture of Group

Question

I'm using regex to capture the dimensions of ads

Source content is an HTML File, and I'm trying to capture for content that looks like:

size[200x400,300x1200] (could be 1-4 different sizes)

I'm trying to an array with the different sizes in it

My capture code looks like this:

$size_declaration = array();
$sizes = array();
$declaration_pattern = "/size\[(\d{2,4}x\d{2,4}|\d{2,4}x\d{2,4},){1,4}\]/";
$sizes_pattern = "/\d{2,4}x\d{2,4}/";

$result = preg_match($declaration_pattern, $html, $size_declaration);
if( $result ) {
    $result = preg_match_all($sizes_pattern, $size_declaration[0], $sizes);
    var_dump($sizes);
}

The code above produces usable results:

$sizes = array(
  [0] => array (
    [0] => '200x400',
    [1] => '300x1200'
  )
)

but it takes quite a bit of code. I was thinking it was possible to collect the results with a single regex, but I couldn't find a result that works. Is there a way to clean this up a bit?

Answer 1

It's not very practical to turn it into a single expression; it would be better to keep them separate; the first expression finds the boundaries and does rudimentary content checks on the inner contents, the second expression breaks it down into individual pieces:

if (preg_match_all('/size\[([\dx,]+)\]/', $html, $matches)) {
    foreach ($matches[0] as $size_declaration) {
        if (preg_match_all('/\d+x\d+/', $size_declaration, $sizes)) {
            print_r($sizes[0]);
        }
    }
}

Answer 2

This one is a little simpler:

$html = "size[200x400,300x600,300x100]";
if (($result = preg_match_all("/(\d{2,4}x\d{2,4}){1,4}/", $html, $matches)) > 0)
    var_dump($matches);
// 
// $matches => 
//     array(
//          (int) 0 => array(
//              (int) 0 => '200x400',
//              (int) 1 => '300x600',
//              (int) 2 => '300x100'
//          ),
//          (int) 1 => array(
//              (int) 0 => '200x400',
//              (int) 1 => '300x600',
//              (int) 2 => '300x100'
//          )
//     )
// 

Answer 3

The only way is to repeat the 4 eventual sizes in the pattern:

$subject = <<<LOD
size[523x800]
size[200x400,300x1200]
size[201x300,352x1200,123x456]
size[142x396,1444x32,143x89,231x456]
LOD;

$pattern = '`size\[(\d{2,4}x\d{2,4})(?:,(\d{2,4}x\d{2,4}))?(?:,(\d{2,4}x\d{2,4}))?(?:,(\d{2,4}x\d{2,4}))?]`';

preg_match_all($pattern, $subject, $matches, PREG_SET_ORDER);
foreach ($matches as &$match) { array_shift($match); }

print_r($matches);

The pattern can also be shorten using references to capture groups:

$pattern = '`size\[(\d{2,4}x\d{2,4})(?:,((?1)))?(?:,((?1)))?(?:,((?1)))?]`';

or with the Oniguruma syntax:

$pattern = '`size\[(\d{2,4}x\d{2,4})(?:,(\g<1>))?(?:,(\g<1>))?(?:,(\g<1>))?]`';