Question

I need a Regex for PHP to do the following:

I want to allow [a-zα-ωá-źа-яա-ֆა-ჰא-ת] and chinese, japanese (more utf-8) letters;
I want to ban [^٩٨٧٦٥٤٣٢١٠۰۱۲۳۴۵۶۷۸۹] (arabic numbers);

This is what i’ve done:

function isValidFirstName($first_name) {
    return preg_match("/^(?=[a-zα-ωá-źа-яա-ֆა-ჰא-ת]+([a-zα-ωá-źа-яա-ֆა-ჰא-ת' -]+)?\z)[a-zα-ωá-źа-яա-ֆა-ჰא-ת' -]+$/i", $first_name);
}

It looks like it works, but if I type letters of more than 1 language, it doesn’t validate.

Examples: Авпа Вапапва á-ź John – doesn’t validate.
John Gger – validates, á-ź á-ź – validates.

I would like to this all of these.

Or if there’s a way, to echo a message if user entered more lingual string.

Answer 1

I can’t reproduce the failure cases here (Авпа Вапапва á-ź John validates just fine), but you can simplify the regex a lot – you don’t need that lookahead assertion:

preg_match('/^[a-zα-ωá-źа-яա-ֆა-ჰא-ת][a-zα-ωá-źа-яա-ֆა-ჰא-ת\' -]*$/i', $first_name)

As far as I can tell from the character ranges you’ve given, you don’t need to exclude the digits because anything outside these character classes will already cause the regex to fail.

Another consideration: If your goal is to allow any letter from any language/script (plus some punctuation and space) you can (if you’re using Unicode strings) further simplify this to:

preg_match('/^\pL[\pL\' -]*$/iu', $first_name)

But generally, I wouldn’t try to validate a name by regular expressions (or any other means): Falsehoods programmers believe about names.