Question

In VB.net I’ve got the following line that removes all non-alphanumeric chars from a string:

return Regex.Replace(build, "[\W]", "")

I now need to extend this to remove non-alphanumeric chars that aren’t [] or _.

I’ve changed the line to:

return Regex.Replace(build, "[\W\[\]_]", "")

However I’m pretty sure that this says

replace non-word or [ or ] or _

how do I negate the tests for the [] and _ chars so that it says

replace non-word and not [ and not ] and not _

Some examples:

"[Foo Bar_123456]" => "[FooBar_123456]"
"[Foo Bar_123-456*]" => "[FooBar_123456]"

(More can be supplied if necessary)

Answer 1

You can use character set subtraction:

[\W-[_\[\]]]

Ie. \W with _, [ and ] removed (latter two needing an escape because they are meta-characters in a character set). A verbatim string @"..." avoids needing to escape each backslash from C#:

var re = new Regex(@"[\W-[_\[\]]]");