PHP: strip_tags - remove only certain tags (and their contents)?

Question

I use the strip_tags() function but I need to remove some tags (and all of their contents).

for example :

<div>
  <p class="test">
    Test A
  </p>
  <span>
    Test B
  </span>
  <div>
    Test C
  </div>
</div>

Let's say, I need to get rid of the P and SPAN tags, and only keep :

<div>
  <div>
    Test C
  </div>
</div>

strip_tags expects as a second parameter the tags that you want to KEEP.

In this particular example I could use striptags($html, "<div>"); but the html I'm scraping and the tags that need to be removed are different all the time.

I searched for hours for a function that suits my needs, but couldn't find anything useful.

Any idea's?

Thorite

Start with DOM and XPath – Eellike

Question already answered here: #9789121 – Alforja

I tried the accepted answer in this post but was not satisfied with the results – Thorite

Answer 1

Use a regular expression. Something like this should work:

$tags = array( 'p', 'span');
$text = preg_replace( '#<(' . implode( '|', $tags) . ')>.*?<\/$1>#s', '', $text);

The demo shows it replacing the desired tags with nothing.

Note that you may need to tweak it more, say, to compensate for whitespace within the tags, or other unknowns that your example does not demonstrate.

Here is the regex to use to capture tags with or without attributes:

'#<(' . implode( '|', $tags) . ')(?:[^>]+)?>.*?<\/$1>#s'

Gaither

@Downvoter - Any comment as to why my functional answer was downvoted? – Gaither

Thanks, this is perfect for my situation. I'm scraping HTML using simple html dom parser and just needed some extra stripping. – Thorite

UPDATE: this regex only strips single tags without attributes... the following seems to work : $text = preg_replace( '#<(' . implode( '|', $tags) . ').*>.*?</\1>#s', '', $text); – Thorite

@Thorite - Your examples don't include attributes - If you wanted to match them, you should be using: '#<(' . implode( '|', $tags) . ')[^>]+>.*?</\1>#s'. – Gaither

@Thorite - Un-accepting this answer on the basis that it didn't fulfill a requirement you never mentioned is ridiculous. It perfectly answers the question you stated, which is proven by the linked demo. – Gaither

I agree, I just didn't want to close this one yet, hoping for some more answers. Your regex now includes tags with attributes, but doesn't work anymore for tags without attributes now. My knowledge of regexes is sadly not well enough to fix this. – Thorite

update: with '#<(' . implode( '|', $tags) . ')[^>]*?>.*?</\1>#s' it seems to work for tags with & without attributes – Thorite

@Thorite - I've updated my answer with a regex to use that will capture tags with or without attributes. – Gaither

Answer 2

You say that you are using Simple HTML DOM (Good! That's the right way to parse HTML). When I need to remove a tag and its contents, I do:

$rows = $html->find("span");

foreach ($rows as $row)
{
  $row->outertext = "";
}

$html->load($html->save());

The last line is required because the DOM gets confused after modifications are made so the entire DOM has to be collapsed and then parsed again so that the changes are made permanent (IMO, a bug in Simple HTML DOM).

The Simple HTML DOM approach is safer and more stable than a regular expression.

Langevin