preg_replace only enclosing tag

Question

I have an html code like below(just a part of it)

<p>
  <strong>
    <div align="center">
      <a onclick="return hs.expand(this)" href="http://example.com/somesome.png">
        <img title="some-bla-bla-text" src="http://example.com/somesome.png" 
             alt="some-bla-bla-text" />
      </a>
    </div>
  </strong><br />
  <strong>
    <div align="center">...

and want to strip it out as

<p>
  <strong>
    <div align="center">
      <img title="some-bla-bla-text" alt="some-bla-bla-text" />
    </div>
  </strong><br />
  <strong>
    <div align="center">...

How can I remove <a onclick="return hs.expand(this)" href="http://example.com/somesome.png"> and its closing tag </a> part of this string?

A regex to match between <a onclick="return hs.expand(this)"....> and </a> would be very helpful I think

possible duplicate of [Best methods to parse HTML with PHP](http://stackoverflow.com/questions/3577641/best-methods-to-parse-html-with-php) — hakre, Jul 29 '11 at 01:32
Related: [Using regex to filter attributes in xpath with php](http://stackoverflow.com/questions/6823032/using-regex-to-filter-attributes-in-xpath-with-php/6823087) — hakre, Jul 29 '11 at 01:34

score 4 · Answer 1 · answered Jul 29 '11 at 01:33

4

Regex isn't powerful enough to do this very well since HTML is not a regular language. It might work in some cases, but then it will be very fragile code that could break when given a different, perfectly valid HTML input. You should look into DOMDocument. It allows you to parse HTML easily.

answered Jul 29 '11 at 01:33

Paul

139,544
27
275
264

That's more a comment than an answer. – hakre Jul 29 '11 at 01:36
2

It answers his question perfectly. He can "`remove ... and it's closing tag `" using `DOMDocument`. – Paul Jul 29 '11 at 01:39
1

@PaulPRO I tried DOMDocument but can't find the right solution, maybe you can 'answer' the question which is better a 'comment' – rasputin Jul 29 '11 at 01:43
@PaulPRO: I'm not objecting that what you wrote is generally right, but it's not answering the question in specific. A simple example code how to do that with DomDocument should not be asked too much, or is it? – hakre Jul 29 '11 at 01:45
@rasputin, since I can't see the code around it, and your anchor tag has no id attribute, I can't come up with a concrete way that will for sure work with all your html unless you want it to remove *all* `` tags in the whole document. – Paul Jul 29 '11 at 01:48
@PaulPRO, yes the actual problem is that, it has no id, name, attribure or something else. I think solution must be done with regex. Maybe someone can give us a regex to match between `` and ` – rasputin Jul 29 '11 at 01:56
@rasputin If you knew the layout of the whole document you could easily navigate to it and remove it. If you know that `` will never change, and there is only one copy of it in the document then you can use regex easily enough. Just do something like `preg_replace('/(.*?)<\/a>/', '$1');` That regex will work in this very specific situation. – Paul Jul 29 '11 at 02:06

score 2 · Answer 2 · answered Jul 29 '11 at 01:37

with some testing and tweaking you might be able to get something like the following to work

$html = preg_replace('/\<a[^>]*\>((?!\<\/a\>).)*\<\/a\>/i', '\1', $html);

it basically says, find an open a tag, then find everything up to the next closing a tag

Alan Moore · Accepted Answer · 2011-07-29T16:29:08.720

1

You can probably do what you want with regexes, but you need to provide more details. Do you want to remove all anchor elements, replacing them with whatever was inside them? Or only those that contain IMG tags? Here's a regex that peels off only those anchor tags whose first attribute is onclick:

$s= preg_replace('~\s*<a\s+onclick="[^"]*"[^>]*>((?:(?!</a>).)*)</a>\s*~is', '$1', $s);

see a demo on ideone.com

EDIT: This regex will match an anchor element with an onclick attribute (not necessarily first).

'~\s*<a[^>]*\s+onclick="[^"]*"[^>]*>((?:(?!</a>).)*)</a>\s*~is'

demo

edited Jul 29 '11 at 16:29

answered Jul 29 '11 at 06:40

Alan Moore

73,866
12
100
156

Yes I wanted "a regex that peels off only those anchor tags whose attribute is onclick(not first but must have)" thanks a lot. Can you make some changes to match these? – rasputin Jul 29 '11 at 12:38

preg_replace only enclosing tag

3 Answers3