Replace multi UTF-8 chars with one same acs ii char by preg_replace()

Question

I have to convert a UTF-8 string to non UTF-8 string, I want to replace:

Ậ,Ẫ,Ẩ,Ầ,Ấ,Â,Ặ,Ẵ,Ẳ,Ằ,Ắ,Ă,Ạ,Ã,Ả,À,Á to A,

Ự,Ữ,Ử,Ừ,Ứ,Ư,Ụ,Ũ,Ủ,Ù,Ú to U,

Ợ,Ỡ,Ở,Ờ,Ớ,Ơ,Ộ,Ỗ,Ổ,Ồ,Ố,Ô,Ọ,Õ,Ỏ,Ò,Ó to O... and much more with same case

By php preg_replace() ?

May I use ?:

$string = preg_replace('/Ậ,Ẫ,Ẩ,Ầ,Ấ,Â,Ặ,Ẵ,Ẳ,Ằ,Ắ,Ă,Ạ,Ã,Ả,À,Á/', 'A', $string);
$string = preg_replace('/Ợ,Ỡ,Ở,Ờ,Ớ,Ơ,Ộ,Ỗ,Ổ,Ồ,Ố,Ô,Ọ,Õ,Ỏ,Ò,Ó/', 'O', $string);
$string = preg_replace('/Ự,Ữ,Ử,Ừ,Ứ,Ư,Ụ,Ũ,Ủ,Ù,Ú/', 'U', $string);

possible duplicate of [Replacing accented characters php](http://stackoverflow.com/questions/3371697/replacing-accented-characters-php) — PleaseStand, Jan 15 '13 at 17:50
no, that posts useless with me, i just want to use `preg_replace()` and not all UTF-8 chars — Jio Cul, Jan 15 '13 at 18:08

score 1 · Answer 1 · answered Jan 15 '13 at 17:55

1

Since regular expressions aren't the optimal way of solving this, may I hint at PHPs iconv facilities:

$string = 'ỬỪỨƯỤ';

// temporarily switch locale
$locale = setlocale(LC_CTYPE, 'en_US.UTF-8');
// use iconv to transliterate
$string = iconv('utf-8', 'us-ascii//TRANSLIT', $string);
// restore locale
setlocale(LC_CTYPE, $locale);

// $string is now "UUUUU"

answered Jan 15 '13 at 17:55

Linus Kleen

33,871
11
91
99

While this is a great idea, it doesn't always come with the expected result. For example, "aÎâa" gets translated as "a^I^aa". It probably depends on other factors as well, since "ỬỪỨƯỤ" for me gets translated as "". – rid Jan 15 '13 at 18:00

score 0 · Accepted Answer · answered Jan 15 '13 at 17:49

You can, if you remove the ,, place everything inside a character class, and add the /u modifier if the input string is UTF-8. Example:

preg_replace('/[ỰỮỬỪỨƯỤŨỦÙÚ]/u', 'U', $string);

You can also use str_replace():

str_replace(array('Ự', 'Ữ', ...), 'U', $string);

or strtr():

strtr($string, 'ỰỮỬ', 'UUU');

score 0 · Answer 3 · answered Jan 15 '13 at 18:11

I've only done this in Java once, but in php the trick will be similar.

In unicode, if you normalize, a diacritic is encoded using combining diacritical marks i.e. one letter and one symbol. Just drop the symbol and you're done.

private static final Pattern DIACRITIC =
        Pattern.compile("[\\p{InCombiningDiacriticalMarks}\\p{IsLm}\\p{IsSk}]+");


public static String replaceCombiningDiacriticalMarks(String text) {
    return DIACRITIC.matcher(Normalizer.normalize(text, Normalizer.Form.NFKD)).replaceAll("");
}

If you also have characters from other alphabets, or mathematical symbols, things get trickier. It's still possible to replace them with pure ascii (like a √ with a v for example), but it gets arbitrary which character to pick.

Replace multi UTF-8 chars with one same acs ii char by preg_replace()

3 Answers3