(PHP 5 >= 5.4.0, PHP 7, PHP 8, PECL intl >= 2.0.0)

Transliterator::transliterate -- transliterator_transliterateTransliterate a string


Object-oriented style

public Transliterator::transliterate(string $string, int $start = 0, int $end = -1): string|false

Procedural style

    Transliterator|string $transliterator,
    string $string,
    int $start = 0,
    int $end = -1
): string|false

Transforms a string or part thereof using an ICU transliterator.



In the procedural version, either a Transliterator or a string from which a Transliterator can be built.


The string to be transformed.


The start index (in UTF-16 code units) from which the string will start to be transformed, inclusive. Indexing starts at 0. The text before will be left as is.


The end index (in UTF-16 code units) until which the string will be transformed, exclusive. Indexing starts at 0. The text after will be left as is.

Return Values

The transformed string on success, or false on failure.


Example #1 Converting escaped UTF-16 code units

= "\u304A\u65E9\u3046\u3054\u3056\u3044\u307E\u3059";
transliterator_transliterate("Hex-Any/Java", $s), "\n";

//now the reverse operation with a supplementary character
$supplChar = html_entity_decode('&#x1D11E;');
mb_strlen($supplChar, "UTF-8"), "\n";
$encSupplChar = transliterator_transliterate("Any-Hex/Java", $supplChar);
//echoes two encoded UTF-16 code units
echo $encSupplChar, "\n";
//and back
echo transliterator_transliterate("Hex-Any/Java", $encSupplChar), "\n";

The above example will output something similar to:


See Also

add a note

User Contributed Notes 5 notes

simonsimcity at gmail dot com
11 years ago
I pretty much like the idea of hdogan, but there's at least one group of characters he's missing: ligature characters.
They're at least used in Norwegian and I read something about French, too ... Some are just used for styling (f.e. fi)

Here's an example that supports all characters (should at least, according to the documentation):
(transliterator_transliterate('Any-Latin; Latin-ASCII; Lower()', "A æ Übérmensch på høyeste nivå! И я люблю PHP! fi"));
// string(41) "a ae ubermensch pa hoyeste niva! i a lublu php! fi"

In this example any character will firstly be converted to a latin character. If that's finished, replace all latin characters by their ASCII replacement.
simonsimcity at gmail dot com
9 years ago
Sorry, for posting it again, but I found a bug in my code:

If you have a character, like the cyrillic ь (a soft-sign - no sound), the "Any-Latin" would translate it to a prime-character, and the "Latin-ASCII" doesn't touch prime-characters. Therefore I added an option to remove all characters, that are higher than \u0100.

Here's my new code, including an example:

var_dump(transliterator_transliterate('Any-Latin; Latin-ASCII; [\u0100-\u7fff] remove',
"A æ Übérmensch på høyeste nivå! И я люблю PHP! есть. fi"));
// string(50) "A ae Ubermensch pa hoyeste niva! I a lublu PHP! est. fi"

Another approach, I found quite helpful (if you by no way want to remove characters ...), try to use iconv() in addition. This surely will just return ASCII characters.


Also an example here:

var_dump(iconv("UTF-8", "ASCII//TRANSLIT//IGNORE", transliterator_transliterate('Any-Latin; Latin-ASCII',
"A æ Übérmensch på høyeste nivå! И я люблю PHP! есть. fi"));
// string(50) "A ae Ubermensch pa hoyeste niva! I a lublu PHP! est'. fi"
hdogan at gmail dot com
11 years ago
You can create slugs easily with:

function slugify($string) {
$string = transliterator_transliterate("Any-Latin; NFD; [:Nonspacing Mark:] Remove; NFC; [:Punctuation:] Remove; Lower();", $string);
$string = preg_replace('/[-\s]+/', '-', $string);
trim($string, '-');

slugify("Я люблю PHP!");
7 years ago
There are some possibly undesirable conversions with ASCII//TRANSLIT//IGNORE or your users may require some custom stuff.

You might want to run a substitution up front for certain things, such as when you want 3 letter ISO codes to replace currency symbols. £ transliterates to "lb", for example, which is incorrect since it's a currency symbol, not a weight symbol (#).

ASCII//TRANSLIT//IGNORE does a great job within the realm of possibility :-)

When it doesn't do something you want it to, you can set up a CSV with one replacement per line and run a function like:

function stripByMap($inputString, $mapFile)
$csv = file($mapFile);
foreach($csv as $line)
$arrLine = explode(',', trim($line));
$inputString = str_replace($arrLine[0],$arrLine[1],$inputString);
return $inputString;

or you can write some regexes. Transliterating using ASCII//TRANSLIT//IGNORE works so well that your map probably won't be very long...
jinmoku at hotmail dot com
13 years ago
OOP version :

= 'àáâãäçèéêëìíîïñòóôõöùúûüýÿ
$rule = 'NFD; [:Nonspacing Mark:] Remove; NFC';

$myTrans = Transliterator::create($rule);

To Top