CODEWITHSUNDEEP
regexperl
user2585000
user2585000

Reputation: 113

perl regex character classes

i'm having trouble regex grouping in perl.

of course this is a much larger problem but it's the same concept i'm dealing with. thank you all in advance for your comments and ideas.

the regex below should only care about this portion of the string to make a decision.

doctor_who:ee

doctor_who:ep

doctor_who:ex

but not

doctor_who:eeh

code:

$str = "doctor_who:ee123ABC451234.123"; #match
$str = "doctor_who:ep123YXZ451234.123"; #match
$str = "doctor_who:ex123451234.123"; #match
$str = "doctor_who:eeh1234LMNOP51234.123";  ##should not match
$str = "doctor_who:abc12341234.123";  ##should not match

$regex = "doctor_who:e[e|p|x]"; #--->problem, what to add/remove?

if ($str =~ m/$regex/){
   print "match!";
}
else {
   print "not matched\n";
}

Upvotes: 0

Views: 191

Answers (3)

stevieb
stevieb

Reputation: 9306

It's trivial with a negative lookahead zero-width assertion. This assumes the only thing you don't want to match specifically is doctor_who:eeh*:

/doctor_who:e(?!eh)[epx]/

In the above example, as long as we match doctor_who:e, we will trigger the lookahead on every match. We can gain efficiency by only using it if absolutely necessary, as noted in the comments by @ikegami:

/doctor_who:e(?:[px]|e(?!h))/

What that'll do is put off the lookahead unless the second char after the : is not p or x, and then only if that char is e.

The second example shown in the comments doesn't use lookarounds at all:

/doctor_who:e(?:[px]|e[^h])/

Upvotes: 2

Craig Estey
Craig Estey

Reputation: 33631

You can do this two ways and you combined the two.

Use a character class:

$regex = 'doctor_who:e[epx]';

Use an alternation:

$regex = 'doctor_who:e(e|p|x)';

These provide the positive match, but they need something else to reject the eeh match.

But, is eeh the only match to reject? That's the only one you mentioned, but, would you prefer something more general, such as any a-z character? (e.g.) What about eec? Should it match or be rejected?

From the examples, ee1 matches. Is that because 1 is not h or because it's a digit?

It isn't totally clear [to me, at least] what the best, most succinct regex should be, because there are some loopholes in the examples. So, here are some regexes based on assumptions I've made as to what you'd really like.

So, if eeh is the only rejection, add:

$regex .= '[^h]';

If you'd like a broader rejection:

$regex .= '[^a-z]';

Or, perhaps, you'd only like to match on numeric:

$regex .= '[0-9]';

Side note: This answer has been edited to reflect the comments below

Upvotes: 1

cberning
cberning

Reputation: 33

Since you are not matching at the end of the string I think you would need two regexes.

$regex = "doctor_who:e[epx]"; # match
$not_regex "doctor_who:e[epx][a-z]"; #-do not match

Then just do

if( $string =~ $regex and $string !~ $not_regex ){}

Upvotes: 1

Related Questions