Improving regex for parsing YouTube / Vimeo URLs

I’ve made a function (in JavaScript) that takes an URL from either YouTube or Vimeo. It figures out the provider and ID for that particular video (demo: http://jsfiddle.net/csjwf/).

function parseVideoURL(url) {

    var provider = url.match(/http://(:?www.)?(w*)/)[2],
        id;

    if(provider == "youtube") {

        id = url.match(/http://(?:www.)?(w*).com/.*v=(w*)/)[2];
    } else if (provider == "vimeo") {

        id = url.match(/http://(?:www.)?(w*).com/(d*)/)[2];
    } else {
        throw new Error("parseVideoURL() takes a YouTube or Vimeo URL");    
    }
    return {
        provider : provider,
        id : id
    }
}

It works, however as a regex Novice, I’m looking for ways to improve it. The input I’m dealing with, typically looks like this:

http://vimeo.com/(id)
http://youtube.com/watch?v=(id)&blahblahblah.....

1) Right now I’m doing three separate matches, would it make sense to try and do everything in one single expression? If so, how?

2) Could the existing matches be more concise? Are they unnecessarily complex? or perhaps insufficient?

3) Are there any YouTube or Vimeo URL’s that would fail being parsed? I’ve tried quite a few and so far it seems to work pretty well.

To summarize: I’m simply looking for ways improve the above function. Any advice is greatly appreciated.

Answer

I am not sure about your question 3), but provided that your induction on the url forms is correct, the regexes can be combined into one as follows:

/http://(?:www.)?(?:(vimeo).com/(.*)|(youtube).com/watch?v=(.*?)&)/

You will get the match under different positions (1st and 2nd matches if vimeo, 3rd and 4th matches if youtube), so you just need to handle that.

Or, if you are quite sure that vimeo’s id only includes numbers, then you can do:

/http://(?:www.)?(vimeo|youtube).com/(?:watch?v=)?(.*?)(?:z|&)/

and the provider and the id will apprear under 1st and 2nd match, respcetively.