Check that string contains non-latin letters

Question

I have the following method to check that string contains only latin symbols.

private boolean containsNonLatin(String val) {
        return val.matches("\\w+");
}

But it returns false if I pass string: my string because it contains space. But I need the method which will check that if string contains letters not in Latin alphabet it should return false and it should return true in all other cases.

Please help to improve my method.

examples of valid strings:

w123.
w, 12
w#123
dsf%&@

If you need to only match ASCII letters and a space, use `return val.matches("[\\p{Alpha} ]+");` — Wiktor Stribiżew, Feb 09 '16 at 08:10
@Wiktor Stribiżew I neeed return false only if I see symbols of another alpabets. Chinese for example — gstackoverflow, Feb 09 '16 at 08:14
`"[\\p{Alpha} ]+"` does not allow Chinese, only ASCII letters and a regular (32 dec. value) space. You can also use `[\\p{L}\\p{M}&&[^\\p{Alpha}]]+` to match one or more any letters but ASCII ones. — Wiktor Stribiżew, Feb 09 '16 at 08:14
`%&@` are not Latin characters. If you want to allow those characters, you'll need a better-defined rule. — shmosel, Feb 09 '16 at 08:21
It returns [true](http://ideone.com/OVfEHG) because it does not contain a letter other than a Latin one. What do you expect? — Wiktor Stribiżew, Feb 09 '16 at 09:03
Please check my answer. I explained what the regex does: it *checks that string contains non-latin letters*. And as far as I understand, if these letters (like `ы`, `ę`) are found, you want to return false. Right? See my fiddles, and feel free to fork them to show what behavior you need. Otherwise, you should re-consider your requirements. — Wiktor Stribiżew, Feb 09 '16 at 09:31

anubhava · Answer 1 · 2016-02-09T08:19:45.867

8

You can use \p{IsLatin} class:

return !(var.matches("[\\p{Punct}\\p{Space}\\p{IsLatin}]+$"));

Java Regex Reference

edited Feb 09 '16 at 08:19

answered Feb 09 '16 at 08:08

anubhava

761,203
64
569
643

I need something like not p{IsLatin} – gstackoverflow Feb 09 '16 at 08:17
You can use `return !(var.matches("[\\p{Space}\\p{IsLatin}]+$"));` to reverse the result. – anubhava Feb 09 '16 at 08:18
Also check updated answer to include punctuation characters in this check. – anubhava Feb 09 '16 at 08:20
java.util.regex.PatternSyntaxException: Unknown character property name {Latin} near index 29 [\p{Punct}\p{Space}\p{IsLatin}]+$ ^ – gstackoverflow Feb 09 '16 at 08:39
[It works in this demo as well](http://ideone.com/jliAQg) What is your Java version? – anubhava Feb 09 '16 at 09:06
I use java 6 version – gstackoverflow Feb 11 '16 at 08:49
1

That's the reason because `\\p{IsLatin}` was added from jdk7 onwards – anubhava Feb 11 '16 at 09:08

Wiktor Stribiżew · Accepted Answer · 2016-02-09T08:25:15.900

I need something like not p{IsLatin}

If you need to match all letters but Latin ASCII letters, you can use

"[\\p{L}\\p{M}&&[^\\p{Alpha}]]+"

The \p{Alpha} POSIX class matches [A-Za-z]. The \p{L} matches any Unicode base letter, \p{M} matches diacritics. When we add &&[^\p{Alpha}] we subtract these [A-Za-z] from all the Unicode letters.

The whole expression means match one or more Unicode letters other than ASCII letters.

To add a space, just add \s:

"[\\s\\p{L}\\p{M}&&[^\\p{Alpha}]]+"

See IDEONE demo:

List<String> strs = Arrays.asList("w123.", "w, 12", "w#123", "dsf%&@", "Двв");
for (String str : strs)
    System.out.println(!str.matches("[\\s\\p{L}\\p{M}&&[^\\p{Alpha}]]+")); // => 4 true, 1 false

score 1 · Answer 3 · answered Feb 09 '16 at 08:11

1

Just add a space to your matcher:

private boolean isLatin(String val) {
    return val.matches("[ \\w]+");
}

answered Feb 09 '16 at 08:11

shmosel

49,289
6
73
138

Bhuwan Prasad Upadhyay · Answer 4 · 2016-02-09T08:15:40.010

0

User this :

public static boolean isNoAlphaNumeric(String s) {
       return s.matches("[\\p{L}\\s]+");
}

\p{L} means any Unicode letter.
\s space character

edited Feb 09 '16 at 08:15

answered Feb 09 '16 at 08:11

Bhuwan Prasad Upadhyay

2,916
1
29
33

Check that string contains non-latin letters

4 Answers4