PHP preg_replace oddity with £ pound sign and ã

Question

Hello

I am applying the following function

<?php

function replaceChar($string){
    $new_string = preg_replace("/[^a-zA-Z0-9\sçéèêëñòóôõöàáâäåìíîïùúûüýÿ]/", "", $string);
    return $new_string;
}

$string = "This is some text and numbers 12345 and symbols !£%^#&$ and foreign letters éèêëñòóôõöàáâäåìíîïùúûüýÿ";

echo replaceChar($string);
?>

which works fine but if I add ã to the preg_replace like

$new_string = preg_replace("/[^a-zA-Z0-9\sçéèêëñòóôõöàáâãäåìíîïùúûüýÿ]/", "", $string);

$string = "This is some text and numbers 12345 and symbols !£%^#&$ and foreign letters éèêëñòóôõöàáâäåìíîïùúûüýÿã";

It conflicts with the pound sign £ and replaces the pound sign with the unidentified question mark in black square.

This is not critical but does anyone know why this is?

Thank you,

Barry

UPDATE: Thank you all. Changed functions adding the u modifier: pt2.php.net/manual/en/… – as suggested by Artefacto and works a treat

function replaceChar($string){
$new_string = preg_replace("/[^a-zA-Z0-9\sçéèêëñòóôõøöàáâãäåìíîïùúûüýÿ]/u", "", $string);
return $new_string;
}

Answer 1

Chances are that your string is UTF-8, but preg_replace() is working on bytes

Answer 2

You might want to take a look at mb_ereg_replace(). As Mark mentioned preg_replace only works on byte level and does not work well with multibyte character encodings.

Cheers,
Fabian

Answer 3

If your string is in UTF-8, you must add the u modifier to the regex. Like this:

function replaceChar($string){
    $new_string = preg_replace("/[^a-zA-Z0-9\sçéèêëñòóôõöàáâäåìíîïùúûüýÿ]/u", "", $string);
    return $new_string;
}

$string = "This is some text and numbers 12345 and symbols !£%^#&$ and foreign letters éèêëñòóôõöàáâäåìíîïùúûüýÿ";

echo replaceChar($string);