Regex for string that do not start with any character of a group and do not contain any of multiple substring

Question

This might be simple, but I have lost a ridiculous amount of hours trying to get there by myself. I need to have a regex that:

1. Do not start with '_' or '-'.
2. Do not end with '_' or '-'.
3. Do not contain '__' or '--' or '_-' or '-_'.

I spent hours on a regex builder and I can't get there. I know I can find a way for each of those criteria separately but I can't have them together.

Update

I forgot to mention it must also only contain these characters [a-zA-Z0-9_-].

Answer 1

Try:

^(?![-_])(?:[a-zA-Z0-9]|[-_](?![-_]))+(?<![_-])$

See: regex101

---

Explanation

• ^ start of string from where...
• (?![-_]) not a - or _ is follwing
• (?: then a repetition of
  • [a-zA-Z0-9] a character that can come at any time
  • | or
  • [-_] a - or _
  • (?![-_]) as long as its not followed by another - or _
• )+
• (?<![_-])$ with the last letter also not beeing - or _

Answer 2

We can use the following pattern:

^[^_-](?!.*(?:__|--|_-|-_))(?:.*[^_-])?$

This pattern says to match:

• ^ from the start of the input
  • [^_-] match any leading character OTHER than underscore or hyphen
  • (?!.*(?:__|--|_-|-_)) assert that __, --, _-, or -_ do not occur
  • (?:.*[^_-])? then match anything so long as it does not end in _ or -
• $ end of the input

Demo

Answer 3

According to your description, here's a regex that should work.

^(?![_-]|.*[_-]{2}|.*[_-]$)

^                 # beginning of the string
(?!...|...|...)   # negative lookahead with OR conditions, neither of the following criterias should be matched
[_-]              # 1. starting with `_` or `-`
.*[_-]{2}         # 2. any consecutive `-` or `_` (`_-`, `-_`, `--` or `__`)
.*[_-]$           # 3. ending with `_` or `-`

See the tests

Answer 4

You can match the strings you are interested in with this regex:

^[^_-](?:[^_-]|[_-][^_-])*$

This matches:

• ^ : beginning of string
• [^_-] : a character which is not an _ or -
• (?:[_-][^_-]|[^_-])* : some number of either:
  • [^_-] : a character which is not an _ or -; or
  • [_-][^_-] : an _ or - which is followed by a character which is not an _ or -;
• $ : end of string

Note that the second part of the alternation requires a character to follow an - or _, this ensures the string cannot end with - or _.

This will be more efficient than using lookaheads.

Demo on regex101

Note that as pointed out by @HaoWu in the comments, the alternation can be simplified to an optional _ or - followed by a character which is not an _ or - i.e.

^[^_-](?:[_-]?[^\n_-])*$

This is slightly more efficient again.

Demo on regex101