Regex - repeating pattern 2

[Borvik]

I'm sure the answer is a fairly simple regex pattern - but I seem to have difficulty with all the nuances of regex.

I'm looking for a regex to validate a repeating pattern of characters. The pattern is AAA-AAA where is any alphanumeric character ([A-Z0-9] - I'll be using strtoupper before the regex). This pattern could repeat any number of times from 1 (AAA) to infinity, but always in 3 alphanumeric character groups separated by a dash(-).

Incomplete patterns should not match (AAA-) or (AAA-AA).

Here is the regex I've come up with so far: /^([A-Z0-9]{3,}-?)*?$/

It appears to work so far, but incomplete patterns that look like (AAA-) in other words a valid set of alphanumeric characters ending with a dash still get a valid result.

Any regex gurus out there that can help with this?

Thanks.

 

[jpadie]

there's a regex way to do this with a lookahead (I think). but the better way to do it would be with a user function

function testString($string){
 $array = explode ('-', $string);
 foreach ($array as $s) {
    if (!preg_match('/^[a-z0-9]{3,3}$/i', $s)) return false;
 }
  return true;
}

 

[jpadie]

and i think that there is a way to do this without lookaheads too

$pattern = '/^[a-z0-9]{3,3}((-[a-z0-9]{3,3})+$|$)/ims';

which says, basically, that
1. match alpha 3 times and only three times
2. then match either the end of the string OR
3. match a repeating string (1 or more times) of a dash followed by three alpha numeric characters followed by either a repeat of the same OR the end of the string.

 

[sen5241b]

The 'i' REGEX modifier will render the search case-insensitive.

 

[jpadie]

exactly. so you don't need to worry about strtoupper()

 

[Borvik]

Duh - case-insensitive modifier. I'm not sure why I didn't see that - it was near the end of the day and my brain was shutting down is the only thing I can come up with.

Thanks guys for the tips guys.

I remembered something about lookaheads this morning and did some digging on that (before I had checked back here) and came up with the following (which I'm adding those modifiers to - now that you mention them):

$pattern = '/^([A-Z0-9]{3,3})(-(?=[A-Z0-9]{3,3})[A-Z0-9]{3,3})*$/ims'

You think that function with the explode and loop would be better jpadie? Any particular reason that would be better than the regex? Would that have better performance or something?

 

[jpadie]

actually i think that my second post of 17h44 will have the best performance. it does not use lookaheads, which will speed up the regex considerably.

 

[Borvik]

It does work - thank you.