CODEWITHSUNDEEP
.netregexregex-lookaroundslookbehind
rory.ap
rory.ap

Reputation: 35270

Regex positive look-ahead not preventing a match

If this regex:

^(?:(?:\([^\(\)]+\))|(?:(?<!\()[^\(\)]+(?!\))))$

matches abc and (abc) but not (abc or abc), why can't I use it in a positive look-ahead like this?

^(?=(?:(?:\([^\(\)]+\))|(?:(?<!\()[^\(\)]+(?!\)))))(?:\(?[a-z]+\)?)$

It matches abc) for example.

Upvotes: 1

Views: 70

Answers (1)

Wiktor Stribiżew
Wiktor Stribiżew

Reputation: 626845

Your first regex can be reduced to ^(?:\([^()]+\)|[^()]+)$. When you used it in the lookahead, you did not anchor it at the end, you did not use $. So, the direct "quick fix" would look like

^(?=(?:\([^()]+\)|[^()]+)$)\(?[a-z]+\)?$

See the regex demo.

The second regex can also be written as mere ^(?:\([a-z]+\)|[a-z]+)$, with two alternatives that either matches a lowercase letter string inside parentheses or without them.

In .NET, you may also use 

^(\()?[a-z]+(?(1)\))$

See demo.

Details

  • ^ - start of string
  • (\()? - an optional capturing group #1 matching a (
  • [a-z]+ - 1+ lowecase letters (\p{Ll}+ matches any lowercase Unicode letters)
  • (?(1)\)) - a conditional construct: if Group 1 matched (if there was an open parenthesis) match )
  • $ - end of string.

Upvotes: 2

Related Questions