Regex : every alphanumeric subdomains, except www

Question

This is a piece of a big regex (php) :

((?:[0-9A-z_-]+))

I use it to valid a subdomain, but I would like to accept every subdomains except "www"

How can I do that when there's an alphanumeric "selector" ?

EDIT : Finally used (\b(?!www\b)(?:[0-9A-Za-z_-]+)), thanks to Tim and ATS

Try `(?!www)`. [More info](http://www.regular-expressions.info/completelines.html) (go to paragraph 'Finding Lines Containing or Not Containing Certain Words') — Laoujin, Feb 19 '14 at 09:45
_Almost_, but not completely, a duplicate of [Regex for ANY string except “www”? (subdomain)](http://stackoverflow.com/q/7099397/2936460). — SQB, Feb 19 '14 at 10:04

score 3 · Answer 1 · answered Feb 19 '14 at 09:46

3

First, that regex isn't correct, it should be ((?:[0-9A-Za-z_-]+)) (or simply ([\w-]+) - the non-capturing group is unnecessary). There are some characters between Z and a that you don't want to match.

Second, use a negative lookahead assertion to make sure that the string you're matching isn't www. To make sure that we're not just taking the submatch ww from www (or fail to match wwwwwhat.sgoing.on), it might be necessary to add word boundary anchors, depending on context:

\b(?!www\b)([\w-]+)

answered Feb 19 '14 at 09:46

Tim Pietzcker

328,213
58
503
561

Works fine, thanks for your explanation, but it accept multiple subdomains (subsubdomain, subsubsubdomain...) and I just one to valid one – Maxime R. Feb 19 '14 at 09:52
Okay I used the following : (\b(?!www\b)(?:[0-9A-Za-z_-]+)) to don't accept dots – Maxime R. Feb 19 '14 at 09:56

score 2 · Accepted Answer · edited Feb 19 '14 at 10:19

2

Try this:

((?!www)(?:[0-9A-Za-z_-]+))

This uses negative lookahead to say "not www".

This assumes that the rest of the regex (as you wrote, it's part of a bigger expression) makes sure that this part contains just a possible subdomain. So this part should match all of a subdomain but shouldn't need to check if it did match all of it.
What you still need to do, that I can't (since I don't know what you're using), is insert the proper boundary detection after www. Probably \b would suffice.

edited Feb 19 '14 at 10:19

Alan Moore

73,866
12
100
156

answered Feb 19 '14 at 09:43

SQB

3,926
2
28
49

Thanks, it worked ! (can only accept your answer in 9 minutes) – Maxime R. Feb 19 '14 at 09:46
No problem, I can wait :) – SQB Feb 19 '14 at 09:47
1

Wont this mean that the domain name cant contain www? What if it was testwwwttest.blah? – Srb1313711 Feb 19 '14 at 09:50
You can replace `[0-9A-Za-z_-]` with simply `[\w-]` and I think you should accept [Tim's answer](http://stackoverflow.com/a/21876644/2936460), not mine. – SQB Feb 19 '14 at 10:01

score 0 · Answer 3 · answered Feb 19 '14 at 09:49

0

This works for me to validate subdomain.

^([a-z0-9]+([\-a-z0-9]*[a-z0-9]+)?\.){0,}([a-z0-9]+([\-a-z0-9]*[a-z0-9]+)?){1,63}(\.[a-z0-9]{2,7})+$

answered Feb 19 '14 at 09:49

Karthick oc

88
1
2
7

1

This doesn't exclude `www` from the match, which is what was asked **explicitly**. – SQB Feb 19 '14 at 10:02

Regex : every alphanumeric subdomains, except www

3 Answers3