PHP preg_match UUID v4

Question

I've got a string that contains UUID v4

$uuid = 'http://domain.com/images/123/b85066fc-248f-4ea9-b13d-0858dbf4efc1_small.jpg';

How would i get the b85066fc-248f-4ea9-b13d-0858dbf4efc1 value from the above using preg_match()?
More info on UUID v4 can be be found here

Anthozoan

not sure but try "^(\{{0,1}([0-9a-fA-F]){8}-([0-9a-fA-F]){4}-([0-9a-fA-F]){4}-([0-9a-fA-F]){4}-([0-9a-fA-F]){12}\}{0,1})$" – Glade

Again, Rufinus, this is wrong, see en.wikipedia.org/wiki/… – Rosmunda

Answer 1

$uuid = 'http://domain.com/images/123/b85066fc-248f-4ea9-b13d-0858dbf4efc1_small.jpg';
preg_match('!/images/\d+/([a-z0-9\-]*)_!i', $uuid, $m);

And

preg_match('/[a-f0-9]{8}\-[a-f0-9]{4}\-4[a-f0-9]{3}\-(8|9|a|b)[a-f0-9]{3‌​}\-[a-f0-9]{12}/', $uuid, $m);

works too. Taken from here, but I don't know if we can rely on that.

Clevis

No, this regex is waay too permissive. – Rosmunda

The question is how to get the uuid from link like that but not from anywhere, e.g. not from text. UUID Definition is here, so I guess my second solution which is, actually, stolen, would work right for this purpose. – Clevis

A more strict regex would be /[a-f0-9]{8}\-[a-f0-9]{4}\-4[a-f0-9]{3}\-(8|9|a|b)[a-f0-9]{3}\-[a-f0-9]{12}/ – Baud

If the regex doesn't work though it looks correct: there are some invisible characters in the code examples which show up in the html source: ‌​ Working version without code highlighting: /[a-f0-9]{8}\-[a-f0-9]{4}\-4[a-f0-9]{3}\-(8|9|a|b)[a-f0-9]{3}\-[a-f0-9]{12}/ – Subtype

@ZachMertes The RFC also allows uppercase letters – Simmie

Answer 2

You can try this simple pattern for uuids

preg_match('/\w{8}-\w{4}-\w{4}-\w{4}-\w{12}/',$uuid,$matches);

Elvia

This is matching ________-____-____-____-____________ – Gabrila

Yeah, so don't use this. At least replace \w with [0-9a-f]: '/[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}/' – Pavis

Answer 3

The accepted answer has several flaws:

• It includes hidden characters that break the regex
• It does not allow for uppercase letters (some or all)
• It does not allow for optional dashes

See Wikipedia for details on allowed representations of a v4 UUID.

A regex check that meets the requirements above is:

preg_match(/(?i)[a-f0-9]{8}[\-]?[a-f0-9]{4}[\-]?4[a-f0-9]{3}[\-]?(8|9|a|b)[a-f0-9]{3}[\-]?[a-f0-9]{12}/, $uuid, $matches);

This option will meet most cases, however a more thorough version would address the following flaws:

• It doesn't allow for the nil UUID (all 0) or the max UUID (all F)
• It doesn't allow for the Windows-based format that includes leading and trailing braces ({...})

A regex check that meets these requirements as well is:

preg_match(/(?i){?[a-f0-9]{8}[\-]?[a-f0-9]{4}[\-]?(4[a-f0-9]{3}|[0]{4}|[F]{4})[\-]?((8|9|a|b)[a-f0-9]{3}|[0]{4}|[F]{4})[\-]?[a-f0-9]{12}}?/, $uuid, $matches);

These pieces mean:

• (?i): force case insensitivity
• {?: 0 or 1 opening braces
• [a-f0-9]{8}: 8 allowed characters
• [\-]?: 0 or 1 dashes
• [a-f0-9]{4}: 4 allowed characters
• [\-]?: 0 or 1 dashes
• (4[a-f0-9]{3}|[0]{4}|[F]{4}): a 4 followed by 3 allowed characters, or 4 zeroes, or 4 Fs
• [\-]?: 0 or 1 dashes
• ((8|9|a|b)[a-f0-9]{3}|[0]{4}|[F]{4}): a 4 followed by 3 allowed characters, or 4 zeroes, or 4 Fs
• [\-]?: 0 or 1 dashes
• [a-f0-9]{12}: 12 allowed characters
• }?: 0 or 1 closing braces

For any of you regex wizards out there, this could further be improved if it's possible to do the following:

• Require that if one dash is present, all four dashes must be present
• Only allow all zeroes or all Fs in some blocks if the entire UUID is all zeroes or Fs
• Only allow an opening or closing brace if both are present

Talanta

Answer 4

When testing if string is matching, then don't forget ^ on beginning and $ on ending in pattern. Without them you can test, if string contains requested string, but not that it match requested string. '/^[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}$/'

Molecular

The question is seeking to extract the UUID from the string, the use of ^ and $ around the pattern for the UUID would not match a string that contained the full URL – Urceolate