natcasesort

w-dot-rosenbach-at-netskill-de (2010-12-14 09:45:44)

Sorting UTF-8 by arbitrary order:

<?php
mb_internal_encoding("UTF-8");

class utf_8_german
{
  // everything else is sorted at the end
  static $order = '0123456789Aa??BbCcDdEeFfGgHhIiJjKkLlMm
NnOo??PpQqRrSs?TtUu?üVvWwXxYyZz';
  static $char2order;
  
  static function cmp($a, $b) {
    if ($a == $b) {
        return 0;
    }
    
    // lazy init mapping
    if (empty(self::$char2order))
    {
      $order = 1;
      $len = mb_strlen(self::$order);
      for ($order=0; $order<$len; ++$order)
      {
        self::$char2order[mb_substr(self::$order, $order, 1)] = $order;
      }
    }
    
    $len_a = mb_strlen($a);
    $len_b = mb_strlen($b);
    $max=min($len_a, $len_b);
    for($i=0; $i<$max; ++$i)
    {
      $char_a= mb_substr($a, $i, 1);
      $char_b= mb_substr($b, $i, 1);
      
      if ($char_a == $char_b) continue;
      $order_a = (isset(self::$char2order[$char_a])) ? self::$char2order[$char_a] : 9999;
      $order_b = (isset(self::$char2order[$char_b])) ? self::$char2order[$char_b] : 9999;
      
      return ($order_a < $order_b) ? -1 : 1;
    }
    return ($len_a < $len_b) ? -1 : 1;
  }
}

// usage example:

$t = array(
  'Birnen', 'Birne', '?pfel', 'Apfel',
);

uasort($t, 'utf_8_german::cmp');
echo '$t: <pre>'.htmlspecialchars(print_r($t,true),null,'UTF-8').'</pre>';
?>

shawn at shawnwilkerson dot com (2009-05-05 10:03:29)

I kept getting varied results using natcasesort and sort on mixed arrays -- per the descriptions.

Sometimes simple is better:

A little snippet of code:

<?php                        if($responders->num_rows) {
                            $i=0;
                            while($row= $responders->fetch_assoc()) {
                                $user=getUserName($row['responderID']);
                                $r[$i]= array("sortname"=>strtolower($user),"userName"=>$user, "userID"=>$row['responderID'], "responderID"=>$row['idresponders']);
                                $i++;
                            }
                            sort($r);
                            print_r($r);                            
                        }

?>

I simply created a lower cased sort field at the front of the result set and then sort by it -- which provides the expected result and leaves the actual needed fields unchanged.

For the curious:  all user information is kept completed in another database (and table) from the content database due to security reasons.  The getUser functions we have written allow us to pull only what is legally allowed without exposing anything else.

This is why a left join or something wasn't used and we have to build a pseudo result array here from both databases.

claude at schlesser dot lu (2009-01-07 00:41:32)

Here a function that will natural sort an array by keys with keys that contain special characters.

<?php
function natksort($array)
{
    $original_keys_arr = array();
    $original_values_arr = array();
    $clean_keys_arr = array();

    $i = 0;
    foreach ($array AS $key => $value)
    {
        $original_keys_arr[$i] = $key;
        $original_values_arr[$i] = $value;
        $clean_keys_arr[$i] = strtr($key, "?????ü?????éèà?", "AOUaouEEAEeeeac");
        $i++;
    }

    natcasesort($clean_keys_arr);

    $result_arr = array();

    foreach ($clean_keys_arr AS $key => $value)
    {
        $original_key = $original_keys_arr[$key];
        $original_value = $original_values_arr[$key];
        $result_arr[$original_key] = $original_value;
    }

    return $result_arr;
}
?>

Hope it will be useful to somebody :)

Kim (2008-12-02 07:45:42)

If your array already is sorted ASC or DESC, then natcasesort, ksort and sort (I only tested those) will invert the sorting order.
This is most troublesome since its impossible to force a sorting order. I had to use array_multisort instead and lose the natrual sorting.
Natcasesort expects a single dimension array, so if your value is an array then it will report a "Array to string conversion" notice per key/value set.

php at method5 dot de (2003-11-06 10:56:56)

vbAlexDOSMan at Yahoo dot com (2003-09-12 01:21:57)

Ulli at Stemmeler dot net: I remade your function -- it's a little more compact now -- Enjoy...
function ignorecasesort(&$array) {
/*Make each element it's lowercase self plus itself*/
/*(e.g. "MyWebSite" would become "mywebsiteMyWebSite"*/
for ($i = 0; $i < sizeof($array); $array[$i] = strtolower($array[$i]).$array[$i], $i++);
/*Sort it -- only the lowercase versions will be used*/
sort($array);
/*Take each array element, cut it in half, and add the latter half to a new array*/
/*(e.g. "mywebsiteMyWebSite" would become "MyWebSite")*/
for ($i = 0; $i < sizeof($array); $i++) {
$this = $array[$i];
$array[$i] = substr($this, (strlen($this)/2), strlen($this));
}
}

dslicer at maine dot rr dot com (2003-06-02 18:41:16)

Something that should probably be documented is the fact that both natsort and natcasesort maintain the key-value associations of the array. If you natsort a numerically indexed array, a for loop will not produce the sorted order; a foreach loop, however, will produce the sorted order, but the indices won't be in numeric order. If you want natsort and natcasesort to break the key-value associations, just use array_values on the sorted array, like so:
natcasesort($arr);
$arr = array_values($arr);

tmiller25 at hotmail dot com (2002-04-25 19:55:25)

add this loop to the function above if you want items which have the same first characters to be listed in a way that the shorter string comes first.
--------------------
/* short before longer (e.g. 'abc' should come before 'abcd') */
for($i=count($array)-1;$i>0;$i--) {
$str_a = $array[$i ];
$str_b = $array[$i-1];
$cmp_a = strtolower(substr($str_a,0,strlen($str_a)));
$cmp_b = strtolower(substr($str_b,0,strlen($str_a)));
if ($cmp_a==$cmp_b && strlen($str_a)<strlen($str_b)) {
$array[$i]=$str_b; $array[$i-1]=$str_a; $i+=2;
}
}
--------------------

Ulli at Stemmeler dot net (2002-04-05 08:01:23)

natcasesort didn't work first time I needed something like this.
Not on my local server, not on my server on the web.
I needed an array sorted ignoring upper and lower cases. In the end lower case array-members stayed at the end of the array.
I replaced it with this function:
---------
function ignorecasesort(&$array) {
$separator="|<>|";
for($i=0;$i<sizeof($array);$i++) { $array[$i]=strtolower($array[$i]).$separator.$array[$i]; }
sort($array);
for($i=0;$i<sizeof($array);$i++) { $this=$array[$i]; $this=explode($separator,$this); $array[$i]=$this[1]; }
}
---------