Regex to match ASCII non-alphanumeric characters

Question

I need a regex to match ASCII non-alphanumeric characters. The regex should not match non-ASCII characters. I am using the following:

   "[\\u0000-\\u002f\\u003a-\\u0040\\u005b-\\u0060\\u007b-\\u007f]"

Can I simplify this regex ?

Are you looking for single ASCII non-alphanumeric character match only? — Braj, Aug 17 '14 at 13:34

score 6 · Accepted Answer · answered Aug 17 '14 at 12:42

6

Yes you can use a character class intersection. Example:

[\\p{ASCII}&&\\P{Alnum}]

This means: intersection between all ascii characters and all non alphanumeric characters

answered Aug 17 '14 at 12:42

Thanks but I need to match ASCII and _non_-alphanumeric characters. So the `"[\\p{ASCII}&&[^\\p{Alnum}]]"` will probably work. – Michael Aug 17 '14 at 12:59
1

@Michael `\P` is the opposite of `\p`. – Unihedron Aug 17 '14 at 13:01
@Michael However since most of the time you'd favour readability, while this regex is shorter, you will find yourself using negated classes in sets more in the future. aka `[a&&[^\s]]` > `[a&&\S]`. – Unihedron Aug 17 '14 at 13:04
Thanks again. You are probably right about using of negated classes. – Michael Aug 17 '14 at 13:07

score 2 · Answer 2 · answered Aug 17 '14 at 12:43

2

You can use this regex in Java

^(?=[^0-9a-zA-Z]+$)\p{ASCII}+$

OR else:

^(?!\p{Alnum}+$)\p{ASCII}+$

answered Aug 17 '14 at 12:43

anubhava

Yes it OP uses them in `matches` method then anchors aren't needed. But if it is used in other methods like `find` then probably needed. – anubhava Aug 17 '14 at 12:47
The OP was talking about matching a single character, which would be `(?!\p{Alnum})\p{ASCII}`. To enforce that condition on the whole string, you would do this: `^(?:(?!\p{Alnum})\p{ASCII})+$`. Your regexes match a string that's all ASCII but not all alphanumeric (that is, it must contain at least one non-alphanumeric character). – Alan Moore Aug 17 '14 at 15:11

score 2 · Answer 3 · edited May 23 '17 at 11:57

2

You can use a set intersection:

"[\\p{ASCII}&&[^\\p{Alnum}]]"

edited May 23 '17 at 11:57

Community

answered Aug 17 '14 at 12:44

Unihedron

Great ! Thank you. Looks like exactly what I need. Too bad it is not a _standard_ regex. – Michael Aug 17 '14 at 13:01
@Michael There is no such thing as a _standard_ regular expression ever since programmers overuse it for everything - Writing the one-liner regex is often more viable than writing an entire finite state machine to parse! – Unihedron Aug 17 '14 at 13:02

3 Answers3