如何从PHP中的UTF8字符“删除变音符号”？

Question

I need to replicate the behavior of MySQL utf8_general_ci collation in PHP. Strictly speaking I need to detect what whould be considered different and what would be considered the same. The case independent part is easy. The problem is utf_general_ci considers characters with diacritics and characters without diacritics to be equal: e = è = é etc.. To replicate that comparison, I'd need to have a way to replace è -> e, é -> e.

The method that comes to my mind is:

echo iconv("utf-8", "ascii//TRANSLIT", "é");

One problem is iconv behaves differently depending on current locale and that's asking for a problem.

The other problem is the input may also contain Cirillic letters that shouldn't be stripped or result in a PHP Notice.

echo iconv("utf-8", "ascii//TRANSLIT", "дом");

Is there a solution or do I have to create manually mapping of each character with diacritic to a one without it?

Answer 1

intl's Transliterator will let you define far more in-depth transliteration rules. The full documentation on transliteration rules can be found on icu-project.org.

$tests = [ "é", "дом" ];

$tl = Transliterator::create('Latin-ASCII;');
foreach($tests as $str) {
    var_dump(
        $tl->transliterate($str)
    );
}

Output:

string(1) "e"
string(6) "дом"