Question

I’m trying to validate YouTube URLs for my application.

So far I have the following:

// Set the youtube URL
$youtube_url = "www.youtube.com/watch?v=vpfzjcCzdtCk";

if (preg_match("/((http\:\/\/){0,}(www\.){0,}(youtube\.com){1} || (youtu\.be){1}(\/watch\?v\=[^\s]){1})/", $youtube_url) == 1)
{
    echo "Valid";
else
{
    echo "Invalid";
}

I wish to validate the following variations of Youtube Urls:

• With and without http://
• With and without www.
• With the URLs youtube.com and youtu.be
• Must have /watch?v=
• Must have the unique video string (In the example above “vpfzjcCzdtCk”)

However, I don’t think I’ve got my logic right, because for some reason it returns true for: www.youtube.co/watch?v=vpfzjcCzdtCk (Notice I’ve written it incorrectly with .co and not .com)

Answer 1

There are a lot of redundancies in this regular expression of yours (and also, the leaning toothpick syndrome). This, though, should produce results:

$rx = '~
  ^(?:https?://)?                           # Optional protocol
   (?:www[.])?                              # Optional sub-domain
   (?:youtube[.]com/watch[?]v=|youtu[.]be/) # Mandatory domain name (w/ query string in .com)
   ([^&]{11})                               # Video id of 11 characters as capture group 1
    ~x';

$has_match = preg_match($rx, $url, $matches);

// if matching succeeded, $matches[1] would contain the video ID

Some notes:

• use the tilde character ~ as delimiter, to avoid LTS
• use [.] instead of \. to improve visual legibility and avoid LTS. (“Special” characters – such as the dot . – have no effect in character classes (within square brackets))
• to make regular expressions more “readable” you can use the x modifier (which has further implications; see the docs on Pattern modifiers), which also allows for comments in regular expressions
• capturing can be suppressed using non-capturing groups: (?: <pattern> ). This makes the expression more efficient.

---

Optionally, to extract values from a (more or less complete) URL, you might want to make use of parse_url():

$url = 'http://youtube.com/watch?v=VIDEOID';
$parts = parse_url($url);
print_r($parts);

Output:

Array
(
    [scheme] => http
    [host] => youtube.com
    [path] => /watch
    [query] => v=VIDEOID
)

Validating the domain name and extracting the video ID is left as an exercise to the reader.

---

I gave in to the comment war below; thanks to Toni Oriol, the regular expression now works on short (youtu.be) URLs as well.