PHP RegExp for nested Div tags

Question

I need a regexp I can use with PHP's preg_match_all() to match out content inside div-tags. The divs look like this:

<div id="t1">Content</div>

I've come up with this regexp so far which matches out all divs with id="t[number]"

/<div id="t(\\d)">(.*?)<\\/div>/

The problem is when the content consists of more divs, nested divs like this:

<div id="t1">Content <div>more stuff</div></div>

Any ideas on how I make my regexp work with nested tags?

Thanks

Answer 1

i think it will be better to use some DOM-instruments

Answer 2

As I recently found out, regex can't do that.

http://stackoverflow.com/questions/1692830/matching-pair-tag-with-regex

I ended up using xpath, and it works like a charm

Answer 3

Don't use regex to parse html.

Answer 4

Try a parser instead:

require_once "simple_html_dom.php";
$text = 'foo <div id="t1">Content <div>more stuff</div></div> bar <div>even more</div> baz  <div id="t2">yes</div>';
$html = str_get_html($text);
foreach($html->find('div') as $e) {
    if(isset($e->attr['id']) && preg_match('/^t\d++/', $e->attr['id'])) {
        echo $e->outertext . "\n";
    }
}

Output:

<div id="t1">Content <div>more stuff</div></div>
<div id="t2">yes</div>

Download the parser here: http://simplehtmldom.sourceforge.net/

Edit: More for my own amusement I tried to do it in regex. Here's what I came up with:

$text = 'foo <div id="t1">Content <div>more stuff</div></div> bar <div>even more</div>
      baz <div id="t2">yes <div>aaa<div>bbb<div>ccc</div>bbb</div>aaa</div> </div>';
if(preg_match_all('#<div\s+id="t\d+">[^<>]*(<div[^>]*>(?:[^<>]*|(?1))*</div>)[^<>]*</div>#si', $text, $matches)) {
    print_r($matches[0]);
}

Output:

Array
(
    [0] => <div id="t1">Content <div>more stuff</div></div>
    [1] => <div id="t2">yes <div>aaa<div>bbb<div>ccc</div>bbb</div>aaa</div> </div>
)

And a small explanation:

<div\s+id="t\d+">  # match an opening 'div' with an id that starts with 't' and some digits
[^<>]*             # match zero or more chars other than '<' and '>'
(                  # open group 1
  <div[^>]*>       #   match an opening 'div'
  (?:              #   open a non-matching group
    [^<>]*         #     match zero or more chars other than '<' and '>'
    |              #     OR
    (?1)           #     recursively match what is defined by group 1
  )*               #   close the non-matching group and repeat it zero or more times
  </div>           #   match a closing 'div'
)                  # close group 1
[^<>]*             # match zero or more chars other than '<' and '>'
</div>             # match a closing 'div'

Now perhaps you understand why people try to persuade you from not using regex for this. As already noted, it will not help if the the html is improperly formed: the regex will make a bigger mess of the output than an html parser, I assure you. Also, the regex will probably make your eyes bleed and your colleagues (or the people who will maintain your software) may come looking for you after seeing what you did. :)

Your best bet is to first clean up your input (using TIDY or similar), and then use a parser to get the info you want.