Java RegEx that matches anything BUT literal string 'NIL' or 'nil'

Question

OK, guys. Here's a Java interview-type question that seems to have stumped some very smart people around here. They actually need this for production code, so it's more than just an interview puzzler.

They need a regular expression, in Java, that returns true if a string literal is anything other than the 3-letter word NIL. The test needs to be case insensitive, and the RegEx itself must do all the work.

So, the RegEx should reject NIL, nil, NiL, nIL, and so on.

It should, however, accept: nile, anil, will, zappa-nil-a, and the empty string.

How many Java developers does it take to write a trivial RegEx? Apparently a lot!

Answer 1

You can do this using a negative lookahead.

With case-insensitive option enabled:

^(?!nil$).*

You could leave off the .* at the end if you don't need to actually return the string in the match. Here is a version without the case-insensitive option:

^(?![nN][iI][lL]$).*

Explanation:

^       # start of string anchor
(?!     # start negative lookahead (fail if...)
   nil    # literal characters 'nil'
   $      # end of string
)       # end lookahead
.*      # consume string (not necessary, but it acts more like a typical regex)

If you want the regex to match nil\n, then use \z instead of $ in the lookahead: ^(?!nil\z).*