Regex to match alphanumeric characters, allowing underscores only in the middle

Question

Right now, I have a regular expression that looks like this:

[A-Za-z][A-Za-z0-9]*

It matches all alphanumeric characters currently. I want it to be able to match with an expression that has a middle underscore, but not if there are consecutive underscores, leading underscores, or trailing underscores. Ie

test__ (not ok)
_test (not ok)
test_ (not ok)
test_string (ok)

What would the regular expression look like?

Like this? `[A-Za-z](?:_[A-Za-z0-9]+)*` Note: Replace the ending * with ? if you only want to allow one underscore — dvo, Nov 01 '19 at 18:27
What about strings that don't have any underscores in between? — CAustin, Nov 01 '19 at 18:32
Consecutive, leading, and trailing underscores should not be matched. — DeviousKoala, Nov 01 '19 at 18:40
@MonkeyZeus - The question seems to be about substring matching, trying to find out. Like can the highlight here be matched ? "\_\_\_\_\_\_\_\_\_`test`\_\_\_\_\_\_" — , Nov 01 '19 at 18:41
@x15 No, that shouldn't be OK because it has consecutive underscores. — DeviousKoala, Nov 01 '19 at 18:42
@MonkeyZeus Yes, those should match because there are no leading or trailing underscores. — DeviousKoala, Nov 01 '19 at 18:44
@ʞᴉɯ Yours appears to be correct, unless there is some obscure case I don't know of. — DeviousKoala, Nov 01 '19 at 18:46
@DeviousKoala - It can't be right since it forces an underscore in the middle and only allows a single one, when it is stated it's optional and can have match many underscores as long as none are consecutive. — , Nov 01 '19 at 18:50
@x15 ^[A-Za-z](?:_?[A-Za-z0-9])* is correct. Post that and I'll give you best answer. — DeviousKoala, Nov 01 '19 at 19:11

ʞᴉɯ · Answer 1 · 2019-11-01T19:02:06.857

1

Try this

 ^[a-zA-Z0-9]+_[a-zA-Z0-9]+$

To make _ optional

^[a-zA-Z0-9]+_{0,1}[a-zA-Z0-9]+$

To allow more than one _

^(?:[a-zA-Z0-9]+_{0,1})+[a-zA-Z0-9]+$

edited Nov 01 '19 at 19:02

answered Nov 01 '19 at 18:48

ʞᴉɯ

5,376
7
52
89

Note that this solution **requires** and a middle underscore which is not the same as **allowing** a middle underscore. This doesn't make the answer wrong, but you should take this in consideration when using this answer. Further more it doesn't allow more then one underscore, so `this_wont_match`. – 3limin4t0r Nov 01 '19 at 18:53
Yeah, this doesn't fit OP's requirements according to the comments on the question, so I'm not sure why this was selected as the answer. – CAustin Nov 01 '19 at 18:55
You guys are right. I just tried it out and it didn't work as expected. I'll still pick a better answer. – DeviousKoala Nov 01 '19 at 18:56
@3limin4t0r so it can do ^[a-zA-Z0-9]+_{0,1}[a-zA-Z0-9]+$ – ʞᴉɯ Nov 01 '19 at 18:56
@DeviousKoala see edits – ʞᴉɯ Nov 01 '19 at 19:06
You could replace `_{0,1}` with `_?` to shorten it a bit. – 3limin4t0r Nov 01 '19 at 19:08

score 0 · Answer 2 · answered Nov 01 '19 at 19:06

0

[A-Za-z][A-Za-z0-9]*(?:_[A-Za-z0-9]+)*

Should meet the requirements. This uses what you already had and adds the option to use an underscore if followed by at least one or more [A-Za-z0-9] characters.

You might want to anchor the solution with ^ and $ to the beginning and ending of the string, but this depends on your use-case.

answered Nov 01 '19 at 19:06

3limin4t0r

19,353
2
31
52

1

Can get slightly better performance with `^(?!\d)[A-Za-z0-9]+(?:_[A-Za-z0-9]+)*$` but I like the approach. +1 – Nov 01 '19 at 20:51
@x15 Hadn't thought of that, I'll leave it as a comment though since I'll have to regenerate the diagram while the result is the same. – 3limin4t0r Nov 01 '19 at 21:00

Regex to match alphanumeric characters, allowing underscores only in the middle

2 Answers2