preg_match_all - regex expression help

Question

I've been given the following regex expression:

 if (preg_match_all("'(http://)?(www[.])?(youtube|vimeo)[^\s]+'is",$prova,$n))
 {
     foreach ($n[3] as $key => $site)
     {
         $video_links[$site][] = $n[0][$key];
     }

However, if I have a string like:

"hello, look at my vimeo video here: http://www.vimeo.com..../ very nice hm?"

Instead of receiving only the url, I'm getting ALSO the word vimeo.

I believe the regex expression is retring more then it should and I would like to retrive ONLY the urls that it finds, not every reference of "vimeo" or "youtube".

Can I request your help in order to narrow the scope of this expression, so that only the URLs are retrieved ?

Try changing the [^\s]+ to \S+ (capital S, which means everything but white-space... — ircmaxell
– ircmaxell, Commented Feb 11, 2011 at 17:54
I don't want to attempt to answer this(as several people have already), but I would like to provide a great site for testing RegEx: gskinner.com/RegExr - it's helped me MANY times! — Sean Walsh
– Sean Walsh, Commented Feb 11, 2011 at 18:40
@s992 - it goes directly to my favourites. Thanks for sharing it. I will try it out more often. — MEM
– MEM, Commented Feb 12, 2011 at 11:17

mario · Accepted Answer · 2011-02-11 18:47:13Z

2

First question mark ? in the regex is unneeded. It makes the preceeding search strings optional, thus also match the bare vimeo word in texts. Try:

preg_match_all("'(http://)(www[.])?(youtube|vimeo)[.][^\s]+'is",

Tip: add (?<![,.)]) at the end if you want to exclude typical interpunction that often screws up such url searches.

As alternative, with http:// and www. optional, but depending on presence of a path:

preg_match_all("'(http://|www[.])*(youtube|vimeo)[.]\w+/[^\s]+'is",

edited Feb 11, 2011 at 18:47

answered Feb 11, 2011 at 17:58

mario

146k20 gold badges243 silver badges293 bronze badges

Sign up to request clarification or add additional context in comments.

6 Comments

netcoder Over a year ago

@mario: Also, [.] is unnecessary. . would do. It was probably meant to be escaped as well (\.). I think you should change that to (www\.)? since the subdomain is optional.

mario Over a year ago

@netcoder: I also do that often. Somehow [.] seems nicer than \. - But the www. should indeed be optional.

ircmaxell Over a year ago

@mario: but a lot of times people will paste urls without the http://, so that part being optional might be wanted...

mario Over a year ago

@ircmaxell: It's not really an url then. But if that's expected content than a wacky (http://|www[.])+ would cover it.

mario Over a year ago

@MEM: See edit. Made it dependend on a path, since you probably don't want plain "www.youtube.com" links without video parameters anyway. But look at TheDogs answer; looks very workable too.

|

The Dog · Accepted Answer · 2011-02-11 18:21:28Z

1

Maybe the following code can help out a bit:

<?php
    //Test string
    $prova = "\"hello, look at my <strong>vimeo</strong> video here:  <a href=\"http://www.vimeo.com..../\" rel=\"nofollow\">http://www.vimeo.com..../</a> very nice hm?\"";
    $prova .= " vimeo vimeo.com/something?id=somethingcrazy&testing=true  ";
    //if we match then capture all matches
    if (preg_match_all("'(http://)?(www\.)?(youtube|vimeo)\.([a-z0-9_/?&+=.]+)'is",$prova,$n)){
        foreach ($n[0] as $key => $site){
            //for each match that matched the whole pattern
            //save the match as a site
            $video_links[$site][] = $n[0][$key];
        }
    }
    //display results
    print_r($video_links);
?>

This will not match the word vimeo. It will match vimeo.com/something?id=somethingcrazy&testing=true and it will match http://www.vimeo.com..../ twice.

answered Feb 11, 2011 at 18:21

The Dog

5052 silver badges11 bronze badges

3 Comments

MEM Over a year ago

Thanks a lot. I've just change $n[0] on foreach argument to $n[3]. So that I can get either youtube or vimeo. PS - If I need to add a third one but within a subdomain, not youtube nor vimeo but, let's say: video.frog.es should the regex expression be: "'(http://)?(www\.)?(youtube|vimeo|video\.frog\.)\.([a-z0-9_/?&+=.]+)'is" ??

The Dog Over a year ago

Assuming a video.frog.es link looks like this: video.frog.es/somevid or video.frog.es/vid It would need to be "'(http://)?(www\.)?(youtube|vimeo|video\.frog)\.([a-z0-9_/?&+=.]+)'is" Your above regex is very similar but you are looking for one too many dots after frog.

MEM Over a year ago

Thank you. I see that with my tried expression I would allow video.frog..es as well, witch is, incorrect. Thanks again. :)

Collectives™ on Stack Overflow

preg_match_all - regex expression help

2 Answers 2

6 Comments

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

6 Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related