(PHP 5 >= 5.4.0, PHP 7, PHP 8, PECL intl >= 2.0.0)
Transliterator::transliterate -- transliterator_transliterate — Transliterate a string
Object-oriented style
Procedural style
$transliterator,$string,$start = 0,$end = -1Transforms a string or part thereof using an ICU transliterator.
transliteratorIn the procedural version, either a Transliterator or a string from which a Transliterator can be built.
stringThe string to be transformed.
startThe 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.
endThe 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.
The transformed string on success, or false on failure.
Example #1 Converting escaped UTF-16 code units
<?php
$s = "\u304A\u65E9\u3046\u3054\u3056\u3044\u307E\u3059";
echo transliterator_transliterate("Hex-Any/Java", $s), "\n";
//now the reverse operation with a supplementary character
$supplChar = html_entity_decode('𝄞');
echo mb_strlen($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", $encSupplChar), "\n";
?>The above example will output something similar to:
お早うございます 1 \uD834\uDD1E 𝄞
I pretty much like the idea of hdogan, but there's at least one group of characters he's missing: ligature characters.
They're at least used in Norwegian and I read something about French, too ... Some are just used for styling (f.e. fi)
Here's an example that supports all characters (should at least, according to the documentation):
<?php
var_dump(transliterator_transliterate('Any-Latin; Latin-ASCII; Lower()', "A æ Übérmensch på høyeste nivå! И я люблю PHP! fi"));
// string(41) "a ae ubermensch pa hoyeste niva! i a lublu php! fi"
?>
In this example any character will firstly be converted to a latin character. If that's finished, replace all latin characters by their ASCII replacement.Sorry, for posting it again, but I found a bug in my code:
If you have a character, like the cyrillic ь (a soft-sign - no sound), the "Any-Latin" would translate it to a prime-character, and the "Latin-ASCII" doesn't touch prime-characters. Therefore I added an option to remove all characters, that are higher than \u0100.
Here's my new code, including an example:
var_dump(transliterator_transliterate('Any-Latin; Latin-ASCII; [\u0100-\u7fff] remove',
"A æ Übérmensch på høyeste nivå! И я люблю PHP! есть. fi"));
// string(50) "A ae Ubermensch pa hoyeste niva! I a lublu PHP! est. fi"
Another approach, I found quite helpful (if you by no way want to remove characters ...), try to use iconv() in addition. This surely will just return ASCII characters.
See: http://stackoverflow.com/a/3542748/517914
Also an example here:
var_dump(iconv("UTF-8", "ASCII//TRANSLIT//IGNORE", transliterator_transliterate('Any-Latin; Latin-ASCII',
"A æ Übérmensch på høyeste nivå! И я люблю PHP! есть. fi"));
// string(50) "A ae Ubermensch pa hoyeste niva! I a lublu PHP! est'. fi"You can create slugs easily with:
<?php
function slugify($string) {
$string = transliterator_transliterate("Any-Latin; NFD; [:Nonspacing Mark:] Remove; NFC; [:Punctuation:] Remove; Lower();", $string);
$string = preg_replace('/[-\s]+/', '-', $string);
return trim($string, '-');
}
echo slugify("Я люблю PHP!");
?>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...