Can it be done with regex?

Question

Having the following regex: ([a-zA-Z0-9//._-]{3,12}[^//._-]) used like pattern="([a-zA-Z0-9/._-]{3,12}[^/._-])" to validate an HTML text input for username, I wonder if is there anyway of telling it to check that the string has only one of the following: ., -, _

By that I mean, that I'm in need of regex that would accomplish the following (if possible)

• alex-how => Valid
• alex-how. => Not valid, because finishing in .
• alex.how => Valid
• alex.how-ha => Not valid, contains already a .
• alex-how_da => Not valid, contains already a -

The problem with my current regex, is that for some reason, accepts any character at the end of the string that is not ._-, and can't figure it out why.

The other problem, is that it doesn't check to see that it contains only of the allowed special characters.

Any ideas?

Answer 1

Try this one out:

^(?!(.*[.|_|-].*){2})(?!.*[.|_|-]$)[a-zA-Z0-9//._-]{3,12}$

Regexpal link. The regex above allow at max one of ., _ or -.

Answer 2

What you want is one or more strings containing all upper, lower and digit characters followed by either one or none of the characters in "-", ".", or "_", followed by at least one character:

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

Answer 3

Hope this will work for you:-

It says starts with characters followed by (-,.,_) and followed and end with characters

^[\w\d]*[-_\.\w\d]*[\w\d]$

Answer 4

Seems to me you want:

^[A-Za-z0-9]+(?:[\._-][A-Za-z0-9]+)?$

Breaking it down:

• ^: beginning of line
• [A-Za-z0-9]+: one or more alphanumeric characters
• (?:[\._-][A-Za-z0-9]+)?: (optional, non-captured) one of your allowed special characters followed by one or more alphanumeric characters
• $: end of line

It's unclear from your question if you wanted one of your special characters (., -, and _) to be optional or required (e.g., zero-or-one versus exactly-one). If you actually wanted to require one such special character, you would just get rid of the ? at the very end.

Here's a demonstration of this regular expression on your example inputs:

http://rubular.com/r/SQ4aKTIEF6

As for the length requirement (between 3 and 12 characters): This might be a cop-out, but personally I would argue that it would make more sense to validate this by just checking the length property directly in JavaScript, rather than over-complicating the regular expression.

Answer 5

^(?=[a-zA-Z0-9/._-]{3,12}$)[a-zA-Z0-9]+(?:[/._-][a-zA-Z0-9]+)?$

or, as a JavaScript regex literal:

/^(?=[a-zA-Z0-9\/._-]{3,12})[a-zA-Z0-9]+(?:[\/._-][a-zA-Z0-9]+)?$/

• The lookahead, (?=[a-zA-Z0-9/._-]{3,12}$), does the overall-length validation.
• Then [a-zA-Z0-9]+ ensures that the name starts with at least one non-separator character.
• If there is a separator, (?:[/._-][a-zA-Z0-9]+)? ensures that there's at least one non-separator following it.

Note that / has no special meaning in a regex. You only have to escape it if you're using a regex literal (because / is the regex delimiter), and you escape it by prefixing with a backslash, not another forward-slash. And inside a character class, you don't need to escape the dot (.) to make it match a literal dot.

Answer 6

The dot in regex has a special meaning: "any character here".

If you mean a literal dot, you should escape it to tell the regex parser so.

Escape dot in a regex range