Desperately Need a Regex

Question

I am trying to validate user id's matching the example:

smith.jack or smith.jack.s

In other words, any number of non-whitespace characters (except dot), followed by exactly one dot, followed by any number of non-whitespace characters (except dot), optionally followed by exactly one dot followed by any number of non-whitespace characters (except dot). I have come up with several variations that work fine except for allowing consecutive dots! For example, the following Regex

^([\S][^.]*[.]{1}[\S][^.]*|[\S][^.]*[.]{1}[\S][^.]*[.]{1}[\S][^.]*)$

matches "smith.jack" and "smith.jack.s" but also matches "smith..jack" "smith..jack.s" ! My gosh, it even likes a dot as a first character. It seems like it would be so simple to code, but it isn't. I am using .NET, btw.

Frustrating.

Answer 1

[^\s.]+\.[^\s.]+(\.[^\s.]+)?

BTW what you asked for allows "." and ".."

Answer 2

that helps?

/^[^\s\.]+(?:\.[^\s\.]+)*$/

or, in extended format, with comments (ruby-style)

/
  ^           # start of line
  [^\s\.]+    # one or more non-space non-dot
  (?:         # non-capturing group
    \.        # dot something
    [^\s\.]+  # one or more non-space non-dot
  )*          # zero or more times
  $           # end of line
/x

you're not clear on how many times you can have dot-something, but you can replace the * with {1,3} or something, to specify how many repetitions are allowed.

i should probably make it clear that the slashes are the literal regex delimiter in ruby (and perl and js, etc).

Answer 3

I think you'd benefit from using + which means "1 or more", instead of * meaning "any number including zero".

Answer 4

You are using the * duplication, which allows for 0 iterations of the given component.

You should be using plus, and putting the final .[^.]+ into a group followed by ? to represent the possibility of an extra set.

Might not have the perfect syntax, but something similar to the following should work.

^[^.\s]+[.][^.\s]+([.][^.\s]+)?$

Or in simple terms, any non-zero number of non-whitespace non-dot characters, followed by a dot, followed by any non-zero number of non-whitespace non-dot characters, optionally followed by a dot, followed by any non-zero number of non-whitespace non-dot characters.

Answer 5

(^.)+|(([^.]+)[.]([^.]+))+

But this would match x.y.z.a.b.c and from your description, I am not sure if this is sufficiently restrictive.

BTW: feel free to modify if I made a silly mistake (I haven't used .NET, but have done plently of regexs)

Answer 6

[^.\s]+\.[^.\s]+(\.([^\s.]+?)?

has unmatched paren. If corrected to

[^.\s]+\.[^.\s]+(\.([^\s.]+?))?

is still too liberal. Matches a.b. as well as a.b.c.d. and .a.b
If corrected to

[^.\s]+\.[^.\s]+(\.([^\s.]+?)?)

doesn't match a.b

Answer 7

^([^.\W]+)\.?([^.\W]+)\.?([^.\W]+)$

This should capture as described, group the parts of the id and stop duplicate periods

Answer 8

^([^.\s]+)\.([^.\s]+)(?:\.([^.\s]+))?$

Answer 9

I'm not familiar with .NET's regexes. This will do what you want in Perl.

/^\w+\.\w+(?:\.\w+)?$/

If .NET doesn't support the non-capturing (?:xxx) syntax, use this instead:

/^\w+\.\w+(\.\w+)?$/

Note: I'm assuming that when you say "non-whitespace, non-dot" you really mean "word characters."

Answer 10

I took a slightly different approach. I figured you really just wanted a string of non-space characters followed by only one dot, but that dot is optional (for the last entry). Then you wanted this repeated.

^([^\s\.]+\.?)+$

Right now, this means you have to have at least one string of characters, e.g. 'smith' to match. You, of course could limit it to only allow one to three repetitions with

^([^\s\.]+\.?){1,3}$

I hope that helps.

Answer 11

I realise this has already been solved, but I find Regexpal extremely helpful for prototyping regex's. The site has a load of simple explanations of the basics and lets you see what matches as you adjust the expression.

Answer 12

RegexBuddy Is a good (non-free) tool for regex stuff