Extract contents out of a website's HTML source using regex

Question

I have the following code that I'm trying to get fixed.

The code:

$pageData = file_get_contents('111234-2.html');
if(preg_match_all('/<a\s+onclick=["\']([^"\']+)["\']/i', $pageData, $links, PREG_PATTERN_ORDER))
     print_r(array_unique($links[1]));
return false;

Some sample HTML where I want it to fetch from:

    <a onclick="doShowCHys=1;ShowWindowN(0,'http://www.example.com/home/Player.aspx?lpk4=116031&amp;playChapter=False',960,540,111234);return false;" href="javascript:void(0);">
<span class="vt">Welcome

        </span>
        <span class="dur">1m 10s</span>
        <span class="" id="bkmimgview-116031">&nbsp;</span>
        <br class="clear">
    </a>

The output I am getting:

Array ( [0] => doShowCHys=1;ShowWindowN(0, )

The output I am hoping for:

Array ( [0] => doShowCHys=1;ShowWindowN(0,'http://www.example.com/home/Player.aspx?lpk4=116031&amp;playChapter=False',960,540,111234);return false;)

How do I achieve this?

Also, you're probably going to see some people make comments about how you shouldn't use regex to pasrse html, and the reason for that is because regular expressions don't handle hierarchical structures like HTML. certain html structures can mess your gegex up — Sam I am says Reinstate Monica, Nov 12 '12 at 17:09

Andrew Cheong · Accepted Answer · 2012-11-12T17:13:59.600

4

You can improve this using a backreference but you're pretty much doomed if there's any more levels of nested quotes.

'/<a\s+onclick=(["\'])((?:(?!\1).)+)\1/i'

The backreference lets you refer to an already-captured group. So, if you caught a " in the first capture, then you want to find a string of non-"s, and likewise, if you caught a ' in the first capture, then you want to find a string of non-'s, and either way end up with that same quote, " or ' respectively.

EDIT:

@vladr offers a much nicer alternative:

'/<a\s+onclick=(["\'])(.*?)\1/i'

Same idea but the non-greedy quantifier makes it unnecessary to test every character for non-whatever-quote-ness. Updated Rubular link: http://rubular.com/r/VXR1nQ4zf5.

edited Nov 12 '12 at 17:13

answered Nov 12 '12 at 17:08

Andrew Cheong

29,362
15
90
145

Here's a Rubular demo: http://rubular.com/r/TAXwfPv0yd. – Andrew Cheong Nov 12 '12 at 17:10
2

or `/ – vladr Nov 12 '12 at 17:10
@vladr - Oh, very nice. Not sure why I didn't think of that. Good points. – Andrew Cheong Nov 12 '12 at 17:11
FWIW, http://www.spaweditor.com/scripts/regex/index.php lets you play with PHP regexps interactively. – vladr Nov 12 '12 at 18:06

Extract contents out of a website's HTML source using regex

1 Answers1