Turkish Character issue in PHP and MySQL

Question

I'm trying to count occurences of all letters in Turkish alphabet in a MySQL database.

When I try to count letter "a" like this, I get correct result :

while($nt=mysql_fetch_array($rt))
{
    $mystring = $nt["word"];

    for($i = 0; $i < strlen($mystring) ; $i++)
    {
        if($mystring[$i] == 'a')
        {
            $a++;
        }
    }
}

When I replace "a", with "ç" I get zero. I already added this code :

$bd = mysql_connect($mysql_hostname, $mysql_user, $mysql_password) or die("database unavailable");
mysql_set_charset('utf8', $bd);

How can I fix my code for Turkish characters? Thanks.

possible duplicate of [UTF-8 all the way through](http://stackoverflow.com/questions/279170/utf-8-all-the-way-through) — vaso123, Dec 02 '14 at 09:05
What [`collation`](https://dev.mysql.com/doc/refman/5.0/en/charset-unicode-sets.html) do you use in DB? — Aleksei Matiushkin, Dec 02 '14 at 09:05
`strlen()` works byte-wise, not character-wise. `ç` in UTF-8 needs more than 1 byte. Use `mb_strlen()` instead. Likewise you can't use `[]` per byte in a string. — AmigoJack, Dec 02 '14 at 09:07
“ç in UTF-8 needs more than 1 byte” is **wrong**. There is U+00E7 which is apparently one-byte symbol. Combining cedilla uses more than one byte, though. — Aleksei Matiushkin, Dec 02 '14 at 09:08
@mudasobwa: no, `U+00E7` is one _codepoint_, but encoded in utf8 it's two _bytes_ : `C3 A7`. — georg, Dec 02 '14 at 09:13
Your script is saved in UTF-8 aswell? `ç` is 0xc3a7 in UTF-8, which are 2 byte. — AmigoJack, Dec 02 '14 at 09:15

georg · Accepted Answer · 2014-12-02T09:22:11.207

In UTF-8 ç is encoded as two bytes (C3 A7), therefore byte-by-byte comparison won't work. Consider substr_count:

$s = "abçdeç";
print substr_count($s, 'ç'); // 2

or use a unicode-aware function like this:

function utf8_char_count($s) {
    $count = [];
    preg_match_all('~.~u', $s, $m);
    foreach($m[0] as $c)
        $count[$c] = isset($count[$c]) ? $count[$c] + 1 : 1;
    return $count;
}

print_r(utf8_char_count('çAüθç')); // [ç] => 2 [A] => 1 [ü] => 1 [θ] => 1

This assumes that your string are actually UTF-8, if this is not the case (hint: var_dump(rawurlencode($str))), check your DB and connection settings (see the linked thread).

Turkish Character issue in PHP and MySQL

1 Answers1