How can I sort an array of UTF-8 strings in PHP?

Question

need help with sorting words by utf-8. For example, we have 5 cities from Belgium.

$array = array(\'Borgloon\',\'Thuin\',\'Lennik\',\'Éghezée\',\'Aubel\');
s         


        

Answer 1

I think you can use strcoll:

setlocale(LC_COLLATE, 'nl_BE.utf8');
$array = array('Borgloon','Thuin','Lennik','Éghezée','Aubel');
usort($array, 'strcoll'); 
print_r($array);

Result:

Array
(
    [0] => Aubel
    [1] => Borgloon
    [2] => Éghezée
    [3] => Lennik
    [4] => Thuin
)

You need the nl_BE.utf8 locale on your system:

fy@Heisenberg:~$ locale -a | grep nl_BE.utf8
nl_BE.utf8

If you are using debian you can use dpkg --reconfigure locales to add locales.

Answer 2

I'd be tempted to loop through the array and convert to English characters before sorting. E.g.

<?php
  $array = array('Borgloon','Thuin','Lennik','Éghezée','Aubel');

  setlocale(LC_CTYPE, 'nl_BE.utf8');

  $newarray = array();
  foreach($array as $k => $v) {
    $newarray[$k] = iconv('UTF-8', 'ASCII//TRANSLIT//IGNORE', $v);
  }

  sort($newarray);
  print_r($newarray);
?>

Probably not the best in terms of processing speed/resources used. But sure does make it easier to understand the code.

Edit:

Thinking about it now, you might be better using some kind of lookup table, something like this:

<?php
  $accentedCharacters = array ( 'à', 'á', 'â', 'ã', 'ä', 'å', 'ç', 'è', 'é', 'ê', 'ë', 'ì', 'í', 'î', 'ï', 'ñ', 'ò', 'ó', 'ô', 'õ', 'ö', 'ø', 'ù', 'ú', 'û', 'ü', 'ý', 'ÿ', 'Š', 'Ž', 'š', 'ž', 'Ÿ', 'À', 'Á', 'Â', 'Ã', 'Ä', 'Å', 'Ç', 'È', 'É', 'Ê', 'Ë', 'Ì', 'Í', 'Î', 'Ï', 'Ñ', 'Ò', 'Ó', 'Ô', 'Õ', 'Ö', 'Ø', 'Ù', 'Ú', 'Û', 'Ü', 'Ý' ); 

  $replacementCharacters = array ( 'a', 'a', 'a', 'a', 'a', 'a', 'c', 'e', 'e', 'e', 'e', 'i', 'i', 'i', 'i', 'n', 'o', 'o', 'o', 'o', 'o', 'o', 'u', 'u', 'u', 'u', 'y', 'y', 'S', 'Z', 's', 'z', 'Y', 'A', 'A', 'A', 'A', 'A', 'A', 'C', 'E', 'E', 'E', 'E', 'I', 'I', 'I', 'I', 'N', 'O', 'O', 'O', 'O', 'O', 'O', 'U', 'U', 'U', 'U', 'Y' );

  $array = array('Borgloon','Thuin','Lennik','Éghezée','Aubel');

  $newarray = array();
  foreach($array as $k => $v) {
    $newarray[$k] = str_replace($accentedCharacters,$replacementCharacters,$v);
  }

  sort($newarray);
  print_r($newarray);
?>

Answer 3

There are great answers here, but this is a dead simple solution for most situations.

function globalsort($array, $in = 'UTF-8', $out = 'ASCII//TRANSLIT//IGNORE')
{
    return usort($array, function ($a, $b) use ($in, $out) {
        $a = @iconv($in, $out, $a);
        $b = @iconv($in, $out, $b);
        return strnatcasecmp($a, $b);
    });
}

And use it like so:

globalsort($array);

Answer 4

As for strcoll I guess it was a nice idea, but doesn't seem to work:

<?php

// Some 
$strings = array('Alpha', 'Älpha', 'Bravo');
// make it German: A, Ä, B
setlocale(LC_COLLATE, 'de_DE.UTF8', 'de.UTF8', 'de_DE.UTF-8', 'de.UTF-8');
usort($strings, 'strcoll');
var_dump($strings);
// as you can see, Ä is last, so this didn't work

A while back I wrote a UTF-8 to ASCII tool that would convert "älph#bla" to "aelph-bla". You could use this to "normalize" your input to make it sortable. It's basically a replacement similar to what @Nick said.

You should use a separate array for sorting, as calling urlify() in a usort() callback would be wasting a lot of resources. try

<?php
// data to sort
$array = array('Borgloon','Thuin','Lennik','Éghezée','Aubel');
// container for modified strings
$_array = array();
foreach ($array as $k => $v) {
    // "normalize" utf8 to ascii
    $_array[$k] = urlify($v);
}
// sort the ASCII stuff (while preserving indexes)
asort($_array);
foreach ($_array as $key => &$v) {
    // copy the original value of the ASCIIfied element
    $v = $array[$k];
}
var_dump($_array);

If you have PHP5.3 or the intl PECL compiled, try @Thai's solution, seems sweet!

Answer 5

If you want to use native solution, so i can propose this one

function compare($a, $b)
{
        $alphabet = 'aąbcćdeęfghijklłmnnoóqprstuvwxyzźż'; // i used polish letters
        $a = mb_strtolower($a);
        $b = mb_strtolower($b);

        for ($i = 0; $i < mb_strlen($a); $i++) {
            if (mb_substr($a, $i, 1) == mb_substr($b, $i, 1)) {
                continue;
            }
            if ($i > mb_strlen($b)) {
                return 1;
            }
            if (mb_strpos($alphabet, mb_substr($a, $i, 1)) > mb_strpos($alphabet, mb_substr($b, $i, 1))) {
                return 1;
            } else {
                return -1;
            }
        }
}

usort($needed_array, 'compare');

Not sure, that is the best solution, but it works for me =)

Answer 6

intl comes bundled with PHP from PHP 5.3 and it only supports UTF-8.

You can use a Collator in this case:

$array = array('Borgloon','Thuin','Lennik','Éghezée','Aubel');
$collator = new Collator('en_US');
$collator->sort($array);
print_r($array);

Output:

Array
(
    [0] => Aubel
    [1] => Borgloon
    [2] => Éghezée
    [3] => Lennik
    [4] => Thuin
)