Match only character set from the same language (like facebook name)?

Question

preg_match(???, 'firstname lastname') // true;
preg_match(???, '서프 누워') // true;
preg_match(???, '서프 lastname') // false;
preg_match(???, '#$@ #$$#') // false;

Currently I use:

'/^([一-龠0-9\s]+|[ぁ-ゔ0-9\s]+|[ก-๙0-9\s]+|[ァ-ヴー0-9\s]+|[a-zA-Z0-9\s]+|[々〆〤0-9\s]+)$/u'

But it only works on some languages.

Answer 1

You need an expression that will match only characters from the same unicode script (and spaces), like:

 ^([\p{SomeScript} ]+|[\p{SomeOtherScript} ]+|...)$

You can build this expression dynamically from the list of scripts:

$scripts = "Hangul Hiragana Han Latin Cyrillic"; // feel free to add more

$re = [];
foreach(explode(' ', $scripts) as $s)
    $re [] = sprintf('[\p{%s} ]+', $s);
$re = "~^(" . implode("|", $re) . ")$~u";

print preg_match($re, 'firstname lastname'); // 1
print preg_match($re, '서프 누워'); // 1
print preg_match($re, '서프 lastname'); // 0
print preg_match($re, '#$@ #$$#'); // 0

Do note however, that it's common for names (at least, in European scripts I'm familiar with) to include characters like dots, dashes and apostrophes, which belong to the "Common" script rather than to a language-specific one. To take these into account, a more realistic version of a "chunk" in the above expression could be like this:

 ((\p{SomeScript}+(\. ?|[ '-]))*\p{SomeScript}+)

which will at least correctly validate L. A. Léon de Saint-Just.

In general, validating people's names is a complicated problem and cannot be solved with 100% accuracy. See this funny post and comments therein for details and examples.