Replace all matches that are not part of HTML code

Question

I have input such as:

<h2 class="role">He played and an important role</h2>

And need to replace the role, but not in the class.

Tricky is, that it may be class="group role something" or so, so I essentially only want to search the real text and not the html, but I need to give back everything.

I'm in PHP and do not have a real good starting point ...

[str_replace](http://php.net/manual/en/function.str-replace.php) — chriz, Feb 28 '13 at 16:20
possible duplicate of [What regex pattern do I need for this?](http://stackoverflow.com/questions/1524377/what-regex-pattern-do-i-need-for-this). The approved answer is all you need. — jeroen, Feb 28 '13 at 16:25
I don't know who `str_replace` might help here. I have tried to isolate the html but did not really figure out how to invert that. — shredding, Feb 28 '13 at 16:25

score 2 · Answer 1 · answered Feb 28 '13 at 16:26

2

Better no preg_ for parsing HTML, use dom:

$input = '<h2 class="role">He played and an important role</h2>';

$dom = new domDocument('1.0', 'utf-8'); 
$dom->loadHTML($input); 
$dom->preserveWhiteSpace = false; 

$element = $dom->getElementsByTagName('h2'); // <--- change tag name as appropriate
$value = $element->item(0)->nodeValue;

// change $value here...

answered Feb 28 '13 at 16:26

michi

6,565
4
33
56

1

I think h2 is just an example HTML, OP wants just any HTML node. – anubhava Feb 28 '13 at 16:27
@anubhava ok, then either change the desired tag (see my edit) or loop through all nodes – michi Feb 28 '13 at 16:30
Problem is, that it's only a snippet, not an entire html document. – shredding Feb 28 '13 at 16:33
@shredding as long as your snipped is not totally screwed html, it should work – michi Feb 28 '13 at 16:36
How'd I loop through all nodes? With xpath? – shredding Feb 28 '13 at 16:37
@shredding welcome! xpath if you have either specific tags or a parent-tag (like ) pls post your solution! – michi Feb 28 '13 at 16:42

MikeM · Accepted Answer · 2013-02-28T17:13:12.353

1

It is better to use the DOM to manipulate HTML, but here is a regex solution.

It will not make the replacement if > appears before < ahead in the string.

$input = '<h2 class="role">He played and an important role</h2>';

$input = preg_replace( '/role(?![^<>]*>[^<>]*(?:<|$))/', 'new role', $input );

echo $input;    
// <h2 class="role">He played and an important new role</h2>

edited Feb 28 '13 at 17:13

answered Feb 28 '13 at 16:50

MikeM

13,156
2
34
47

That's cool, but it does not work for `
He played and an important role
`. – shredding Feb 28 '13 at 16:56
Hm, not in my PHP Version (5.4.4) – shredding Feb 28 '13 at 17:05
1

@shredding. Updated answer following your new example input. – MikeM Feb 28 '13 at 17:14
Okay, this is just one line over multipe lines with recursive calling of methods and looping. I'll take that. – shredding Feb 28 '13 at 17:20

Replace all matches that are not part of HTML code

2 Answers2

He played and an important role