Regular Expression special characters

Question

I have this code with this regular expression:

string lPattern = "\\S*[a-zA-Z]+\\S*";

System.Text.RegularExpressions.Regex lRegex =
    new System.Text.RegularExpressions.Regex(lPattern);

if (!lRegex.IsMatch(phone))
{
    return true;
}

I want to make it so if space x space " x " is in the string is will also return true. How would I change this regular expression to include this? Thanks!

Can you give some specific examples of strings you want to match, and just as important, what you want to *not* match. — Mark Byers, Apr 25 '12 at 18:32
don't you mean white space `\s` instead of "everything except white space" `\S` - what is the actual specification? — Aprillion, Apr 25 '12 at 18:37
I'm considering something that is only digits a valid phone. But sometimes I will have a " x " that will also be valid. — , Apr 25 '12 at 18:53
@deathApril: It has to do with digits, because the OP is doing negative invalidation. If it matches the regex, the phone number is invalid. But there are problems with this, as per my answer below. — Mike Ryan, Apr 25 '12 at 19:04
@MikeRyan i see now - maybe http://stackoverflow.com/q/123559/1176601 will help? — Aprillion, Apr 25 '12 at 19:10
@deathApril: It's certainly a lot to think about! Given an arbitrary international #, the accepted answer to that question is probably going to produce ambiguous cases. But the huge discussion about how hard it is to validate a number should give anyone pause. We'll let the OP determine whether or not just doing a few specific negative invalidations is enough or what route works for him. — Mike Ryan, Apr 25 '12 at 19:25

score 2 · Answer 1 · answered Apr 25 '12 at 18:32

2

" x |\\S*[a-zA-Z]+\\S*"

or am i missing something?

answered Apr 25 '12 at 18:32

Aprillion

21,510
5
55
89

@BrokenGlass no, i don't.. but OP might - just asked him/her in a comment under the question :) – Aprillion Apr 25 '12 at 18:34

score 2 · Accepted Answer · edited May 23 '17 at 10:24

Apparently, this code is trying to validate a phone number by matching any alpha characters in the phone number -- if a match is made, then the phone number is invalid.

So you want any alpha character (optionally embedded inside non-whitespace characters, though these are listed as optional, so they're actually irrelevant) except for the sequence x (space x space) to match and trigger failure.

This sounds like you want accept the case of people embedding an extension into the phone number, as in: 123-123-1234 x 789

.*([a-zA-Z])+(?<! x).*

comes very close.

EDIT: This uses a negative look-behind construction. (?<! allows the regex to keep going only if from the current match position, the previous characters do not match what's inside the remainder of the parentheses. So, if there was a leading x it would fail the match.

--

Using this regex, the following don't match:

123-123-1234
123-123-1234 x789
123-123-1234 x 789

But it does match:

123-123a-1234
123-123 asdfasdf 12312
123-123x-123123
123-123 xfasdf

And the ones that match are invalid, by virtue of the ! in your if-statement.

But having said that, your code would be much easier if you wrote the phone number validation as a positive validation (i.e. what do you allow), not a negative invalidation (what do you specifically avoid). By writing what you allow, you'll be able to prevent edge cases much easier.

For example, the following is not matched by both regex's:

123-###-1234
123-!$#&^(##^#*@#(!@&#(#-1234

and this would return true from your function. Is that what you really want?

I strongly suggest positive validation, not negative invalidation.

EDIT: @Aprillion suggests this previous linked question: A comprehensive regex for phone number validation

This shows how complicated phone number validation can be (especially if you're validating for more than 1 country). There's a lot of good suggestions in there, though it's not be the be-all-end-all summary. My inclination would be on the front-end to allow only digits (w/ an optional free-from comment field) for phone number entry.

I must admit also, that, it does give me a little less confidence on positive validation -- especially if you have to deal w/ an arbitrary international #.

A lot would be dependent upon the OP's real requirements.

The RegEx you provided does what I need. Doesn't handle everything, but enough for what I am doing. Thanks! — , Apr 25 '12 at 20:32
Cool. I added a quick line about negative look-behind for a little more about the regex. — Mike Ryan, Apr 25 '12 at 20:38

score 0 · Answer 3 · answered Apr 25 '12 at 18:39

0

string lPattern = @"^\s*[a-zA-Z]+\s*$";

// Start of string
// Zero or more white space
// Letters any case 1 or more
// Zero or more white space
// End of string

\S means anything other than whitespace
\s is to match for whitespace

answered Apr 25 '12 at 18:39

James Kyburz

13,775
1
32
33

score 0 · Answer 4 · answered Apr 25 '12 at 19:25

If you want to validate a phone number, I would also recommend turning the logic around - i.e. try to find a valid phone number rather than invalid. It's easier to restrict what you want to match than to think of all possibilities that you want to reject.

I would use this regex:

(\d+(?:-|\s)?)+(?:\s*x\s*\d+)?

And the code would become:

string lPattern = @"(\d+(?:-|\s)?)+(?:\s*x\s*\d+)?";

        System.Text.RegularExpressions.Regex lRegex = new System.Text.RegularExpressions.Regex(lPattern);

        if (lRegex.IsMatch(phone)) return true;

If that's not what you are after then you'll have to provide some context.

Regular Expression special characters

4 Answers4