Multibyte Array Sortierung

Chriz · 15.01.2009, 03:25

Hallo,

ich habe ein Sortierskript für Multibyte-Arrays geschrieben, um endlich deutsche Umlaute unter UTF-8 gescheit sortiert zu bekommen.

Ich würde mich freuen wenn ihr es mal testen oder mir Performance-Verbesserungen vorschlagen würdet.

Gern auch Hinweise auf interne PHP-Funktionen die schon genau das können (sort und strcoll sortieren scheinbar noch strikt nach Zeichentabelleneintrag).

Chriz · 15.01.2009, 03:27

Ja ich würds gern posten, aber ein Bug im Forum schneidet alles bei den UTF-8 Zeichen ab.

Hier die base64 Version:

PHP-Code:

  <?php
echo base64_decode(<<<B64
<?php
error_reporting(E_ALL | E_STRICT);
header("Content-type: text/html; charset=utf-8");

/**
 * @desc
 * Sort arrays with multibyte values using a predefined or custom alphabet.
 *
 * @author Christian Reinecke <reinecke@bajoodoo.de>
 * @version 1.0
 * @since 2008-01-15
 *
 * @example
 * <code>
 *     // quick
 *     $array = array("Zebra", "Engländer", "Englander", "Chinese", "Äffchen", "Überseeboot",
 *                    "Unterseeboot", "ß", "Sz", "Bürger", "Burger", "burger", "bürger", "ÄÖÜ"));
 *     AIS_Util_MultibyteSort::staticSort($array);
 *     // Äffchen, ÄÖÜ, Burger, Bürger, burger, bürger, Chinese, Englander, Engländer, Sz, ß,
 *     // Unterseeboot, Überseeboot, Zebra
 *
 *
 *     // with custom setup
 *     $mbsort = new AIS_Util_MultibyteSort(AIS_Util_MultibyteSort::AlphabetGerman,
 *                                          AIS_Util_MultibyteSort::EncodingUtf8,
 *                                          AIS_Util_MultibyteSort::CasePreferenceUpper);
 *     $mbsort->sort($array);
 *     // same result as above
 *     $mbsort->sort($otherArray);
 * </code>
 *
 * @todo
 * This class can not handle unknown special characters (fallback to default behaviour),
 * which makes it desirable to find a global, sorted alphabet. Feel free to add more char maps.
 */
class AIS_Util_MultibyteSort
{
    const EncodingUtf8    = "UTF-8";
    const EncodingDefault = self::EncodingUtf8;

    const CasePreferenceUpper   = true;
    const CasePreferenceLower   = false;
    const CasePreferenceDefault = self::CasePreferenceUpper;

    const AlphabetGerman  = "german";
    const AlphabetDefault = self::AlphabetGerman;

    /**
     * predefined loose alphabets
     */
    protected static $alphabets = array(
        self::AlphabetGerman => array("A" => "Ä", "a" => "äà", "O" => "Ö", "o" => "ö", "U" => "Ü",
                                      "u" => "ü", "s" => "ß"));

    /**
     * multibyte encoding
     * @see self::EncodingDefault
     * @var string
     */
    protected $encoding;

    /**
     * used loose alphabet table
     * @see self::$alphabets
     * @var array
     */
    protected $alphabetLoose;

    /**
     * array map (single special char => ASCII-128)
     * @var array
     */
    protected $alphabet;

    /**
     * case flag, whether upper case chars are preferred to lower case chars
     * @see self::CasePreferenceDefault
     * @var bool
     */
    protected $case;

    /**
     * @param array|string [optional] loose alphabet or key for predefined loose alphabet
     * @param string       [optional] encoding for multibyte functions
     * @param bool         [optional] prefer uppercase char to lower-case (Aa or aA)
     */
    public function __construct($alphabet = self::AlphabetDefault, $encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault)
    {
        $this->setEncoding($encoding);
        $this->setAlphabet($alphabet);
        $this->setUpperCasePreference($preferUpperCase);
    }

    /**
     * @desc shortcut
     */
    public static function staticSort(&$array, $alphabet = self::AlphabetDefault, $encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault)
    {
        $self = new self($alphabet, $encoding, $preferUpperCase);
        $self->sort($array);
    }

    /**
     * @desc sort array with the constructor-given settings
     */
    public function sort(&$array)
    {
        /**
         * Got this message: "Warning: usort() [function.usort]: Invalid comparison function"?
         * Then do not call this method from a static context or change your error_reporting level
         * to E_ALL | E_STRICT.
         */
        usort($array, array($this, "sortString"));
    }

    /**
     * @desc multibyte string compare (something like we would expect behind mb_strcmp())
     * @return int sort order value (1, 0, -1)
     */
    protected function sortString($a, $b)
    {
        $ax = mb_strlen($a, $this->encoding);
        $bx = mb_strlen($b, $this->encoding);
        for ($i = 0, $x = min($ax, $bx); $i < $x; ++$i) {
            $result = $this->charCmp(mb_substr($a, $i, 1, $this->encoding),
                                     mb_substr($b, $i, 1, $this->encoding));
            if ($result != 0) {
                return $result;
            }
        }
        return $this->intCmp($ax, $bx);
    }

    /**
     * @desc integer compare
     * @return int sort order value (-1, 0, 1)
     */
    protected function intCmp($a, $b)
    {
        return ($a == $b) ? 0 : ($a < $b ? -1 : 1);
    }

    /**
     * @desc multibyte char compare
     * @return int sort order value (-1, 0, 1)
     */
    protected function charCmp($a, $b)
    {
        // check if characters are known as special chars
        $ai = isset($this->alphabet[$a]); // ai = a isset
        $bi = isset($this->alphabet[$b]);

        if ($ai && $bi) {
            // both are known special chars
            $ar = $this->alphabet[$a]; // ar = a representation (ASCII-128)
            $br = $this->alphabet[$b];
            $result = $this->charCaseCmp($ar, $br);
            if ($result == 0 && $a != $b) {
                // they aren't equal, but their representation is, so check position in original array
                $ap = mb_strpos($this->alphabetLoose[$ar], $a, 0, $this->encoding);
                $bp = mb_strpos($this->alphabetLoose[$br], $b, 0, $this->encoding);
                $result = $this->intCmp($ap, $bp);
            }
        } else if ($ai) {
            // $a is a known special char, $b not
            $result = $this->charCaseCmp($this->alphabet[$a], $b);
             // so they are not equal, $result = 0 means $b is "smaller"
            $result = ($result == 0) ? 1 : $result;
        } else if ($bi) {
            // $b is a known special char, $a not
            $result = $this->charCaseCmp($a, $this->alphabet[$b]);
            // so they are not equal; $result = 0 means $a is "smaller"
            $result = ($result == 0) ? -1 : $result;
        } else {
            // both are unknown characters
            $result = $this->charCaseCmp($a, $b);
        }
        return $result;
    }

    /**
     * @desc multibyte char case compare
     * @return sort order value (-1, 0, 1)
     */
    protected function charCaseCmp($a, $b)
    {
        if ($a == $b) {
            // they are equal, no check required
            return 0;
        }
        $A = mb_strtoupper($a, $this->encoding);
        $B = mb_strtoupper($b, $this->encoding);
        $result = strcmp($A, $B);
        if ($result == 0) {
            // their mb_strtoupper() value is equal, select compare value depending on case preference
            $result = ($A != $a) ? $this->case : ($this->case * -1);
        }
        return $result;
    }

    protected function setAlphabet($alphabet)
    {
        if (!is_array($alphabet)) {
            // use $alphabet as key for predefined alphabets
            $alphabet = self::$alphabets[$alphabet];
        }
        // store loose alphabet to differ between special chars with the same ASCII-128 representation
        $this->alphabetLoose = $alphabet;
        foreach ($this->alphabetLoose as $order => $char) {
            for ($i = 0, $x = mb_strlen($char, $this->encoding); $i < $x; ++$i) {
                // use each multibyte char as key with its ASCII-128 representation as key
                $this->alphabet[mb_substr($char, $i, 1, $this->encoding)] = $order;
            }
        }
    }

    protected function setEncoding($encoding)
    {
        $this->encoding = $encoding;
    }

    protected function setUpperCasePreference($preferUpperCase)
    {
        $this->case = $preferUpperCase ? 1 : -1;
    }
}
B64
);
?>

nikosch · 15.01.2009, 12:27

Cool. Ein Geheimposting!

PS: Schöne Sache. Ich würde noch case-insensitivity anregen.

Chriz · 15.01.2009, 13:54

PHP-Code:

  <?php
/**
 * @desc
 * Sort arrays with multibyte values using a predefined or custom alphabet.
 *
 * @author Christian Reinecke <reinecke@bajoodoo.de>
 * @version 1.0
 * @since 2008-01-15
 *
 * @todo
 * This class can not handle unknown special characters (fallback to default behaviour),
 * which makes it desirable to find a global, sorted alphabet. Feel free to add more char maps.
 */
class AIS_Util_MultibyteSort
{
    const EncodingUtf8    = "UTF-8";
    const EncodingDefault = self::EncodingUtf8;

    const CasePreferenceUpper   = true;
    const CasePreferenceLower   = false;
    const CasePreferenceDefault = self::CasePreferenceUpper;

    /**
     * multibyte encoding
     * @see self::EncodingDefault
     * @var string
     */
    protected $encoding;

    /**
     * used loose alphabet table
     * @see self::$alphabets
     * @var array
     */
    protected $alphabetLoose;

    /**
     * array map (single special char => ASCII-128)
     * @var array
     */
    protected $alphabet;

    /**
     * case flag, whether upper case chars are preferred to lower case chars
     * @see self::CasePreferenceDefault
     * @var bool
     */
    protected $case;

    /**
     * @param array|string [optional] loose alphabet or key for predefined loose alphabet
     * @param string       [optional] encoding for multibyte functions
     * @param bool         [optional] prefer uppercase char to lower-case (Aa or aA)
     */
    public function __construct($alphabet, $encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault)
    {
        $this->setEncoding($encoding);
        $this->setAlphabet($alphabet);
        $this->setUpperCasePreference($preferUpperCase);
    }

    /**
     * @desc shortcut
     */
    public static function staticSort(&$array, $alphabet, $encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault)
    {
        $self = new self($alphabet, $encoding, $preferUpperCase);
        $self->sort($array);
    }

    /**
     * @desc sort array with the constructor-given settings
     */
    public function sort(&$array)
    {
        /**
         * Got this message: "Warning: usort() [function.usort]: Invalid comparison function"?
         * Then do not call this method from a static context or change your error_reporting level
         * to E_ALL | E_STRICT.
         */
        usort($array, array($this, "sortString"));
    }

    /**
     * @desc multibyte string compare (something like we would expect behind mb_strcmp())
     * @return int sort order value (1, 0, -1)
     */
    protected function sortString($a, $b)
    {
        $ax = mb_strlen($a, $this->encoding);
        $bx = mb_strlen($b, $this->encoding);
        for ($i = 0, $x = min($ax, $bx); $i < $x; ++$i) {
            $result = $this->charCmp(mb_substr($a, $i, 1, $this->encoding),
                                     mb_substr($b, $i, 1, $this->encoding));
            if ($result != 0) {
                return $result;
            }
        }
        return $this->intCmp($ax, $bx);
    }

    /**
     * @desc integer compare
     * @return int sort order value (-1, 0, 1)
     */
    protected function intCmp($a, $b)
    {
        return ($a == $b) ? 0 : ($a < $b ? -1 : 1);
    }

    /**
     * @desc multibyte char compare
     * @return int sort order value (-1, 0, 1)
     */
    protected function charCmp($a, $b)
    {
        // check if characters are known as special chars
        $ai = isset($this->alphabet[$a]); // ai = a isset
        $bi = isset($this->alphabet[$b]);

        if ($ai && $bi) {
            // both are known special chars
            $ar = $this->alphabet[$a]; // ar = a representation (ASCII-128)
            $br = $this->alphabet[$b];
            $result = $this->charCaseCmp($ar, $br);
            if ($result == 0 && $a != $b) {
                // they aren't equal, but their representation is, so check position in original array
                $ap = mb_strpos($this->alphabetLoose[$ar], $a, 0, $this->encoding);
                $bp = mb_strpos($this->alphabetLoose[$br], $b, 0, $this->encoding);
                $result = $this->intCmp($ap, $bp);
            }
        } else if ($ai) {
            // $a is a known special char, $b not
            $result = $this->charCaseCmp($this->alphabet[$a], $b);
             // so they are not equal, $result = 0 means $b is "smaller"
            $result = ($result == 0) ? 1 : $result;
        } else if ($bi) {
            // $b is a known special char, $a not
            $result = $this->charCaseCmp($a, $this->alphabet[$b]);
            // so they are not equal; $result = 0 means $a is "smaller"
            $result = ($result == 0) ? -1 : $result;
        } else {
            // both are unknown characters
            $result = $this->charCaseCmp($a, $b);
        }
        return $result;
    }

    /**
     * @desc multibyte char case compare
     * @return sort order value (-1, 0, 1)
     */
    protected function charCaseCmp($a, $b)
    {
        if ($a == $b) {
            // they are equal, no check required
            return 0;
        }
        $A = mb_strtoupper($a, $this->encoding);
        $B = mb_strtoupper($b, $this->encoding);
        $result = strcmp($A, $B);
        if ($result == 0) {
            // their mb_strtoupper() value is equal, select compare value depending on case preference
            $result = ($A != $a) ? $this->case : ($this->case * -1);
        }
        return $result;
    }

    protected function setAlphabet($alphabet)
    {
        if (!is_array($alphabet)) {
            // use $alphabet as key for predefined alphabets
            $alphabet = self::$alphabets[$alphabet];
        }
        // store loose alphabet to differ between special chars with the same ASCII-128 representation
        $this->alphabetLoose = $alphabet;
        foreach ($this->alphabetLoose as $order => $char) {
            for ($i = 0, $x = mb_strlen($char, $this->encoding); $i < $x; ++$i) {
                // use each multibyte char as key with its ASCII-128 representation as key
                $this->alphabet[mb_substr($char, $i, 1, $this->encoding)] = $order;
            }
        }
    }

    protected function setEncoding($encoding)
    {
        $this->encoding = $encoding;
    }

    protected function setUpperCasePreference($preferUpperCase)
    {
        $this->case = $preferUpperCase ? 1 : -1;
    }
}
?>

Chriz · 15.01.2009, 14:04

Gibts doch nicht, ich kann das Anwendungsbeispiel nichtmal eingetippt hier posten

OK dann zur Erklärung, ein $alphabet ist ein assoziativer Array, der Schlüssel muss ein ASCII-128-Zeichen sein, dessen Werte die Sonderzeichen, die bei der Sortierung dann direkt nach diesem ASCII-128-Zeichen einsortiert werden. Möglich ist als Wert mehr als ein Zeichen, also ä oder Akcent a (das kann ich hier offenbar nicht posten).

nikosch · 15.01.2009, 14:22

Yepp, bekanntes Problem: http://www.php.de/board-support/4935...uckt-text.html (Zeichensatz-Probleme, Board verschluckt Text)

Chriz · 02.03.2009, 20:13

Hier nochmal die verständliche Version mit erweitertem Kommentar und Beispiel, ich hoffe das klappt jetzt - habe das Akcent a entfernt. Die API hat sich etwas verändert, das Alphabet muss jetzt nicht mehr übergeben werden, sondern ist vordefiniert.

PHP-Code:

  <?php
/**
 * @desc
 * Sort arrays with multibyte values using a predefined or custom alphabet.
 *
 * @author Christian Reinecke <reinecke@bajoodoo.de>
 * @version 1.1
 * @since 2009-01-15
 *
 * @example
 * <code>
 *     // quick
 *     $array = array("Zebra", "Engländer", "England", "Chinese", "Äffchen", "Überseeboot",
 *                    "Unterseeboot", "ß", "Sz", "Bürger", "Burger", "burger", "bürger", "ÄÖÜ");
 *     AIS_Util_MultibyteSort::staticSort($array);
 *     print_r($array);
 *     // Array (
 *     // [0] => Äffchen [1] => ÄÖÜ [2] => Burger [3] => Bürger [4] => burger [5] => bürger
 *     // [6] => Chinese [7] => England [8] => Engländer [9] => Sz [10] => ß [11] => Unterseeboot
 *     // [12] => Überseeboot [13] => Zebra
 *     // )
 *
 *
 *     // with custom setup
 *     $mbsort = new AIS_Util_MultibyteSort("UTF-8", true);
 *     $mbsort->sort($array);
 *     print_r($array)
 *     // same result as above
 *     $mbsort->sort($otherArray);
 * </code>
 *
 * @todo
 * This class can not handle unknown special characters (fallback to default behaviour),
 * which makes it desirable to find a global, sorted alphabet. Feel free to add more special chars.
 *
 * Don't forget to send the correct charset in your header
 * <code>
 *     header("Content-type: text/html; charset=utf-8");
 * </code>
 */
class AIS_Util_MultibyteSort
{
    const EncodingUtf8    = "UTF-8";
    const EncodingDefault = self::EncodingUtf8;

    const CasePreferenceUpper   = true;
    const CasePreferenceLower   = false;
    const CasePreferenceDefault = self::CasePreferenceUpper;

    protected static $instances = array();
    protected static $alphabetLoose = array("A" => "Ä", "a" => "äæå", "O" => "Ö", "o" => "ö", "U" => "Ü",
                                            "u" => "ü", "s" => "ß");

    /**
     * multibyte encoding
     * @see self::EncodingDefault
     * @var string
     */
    protected $encoding;

    /**
     * array map (single special char => ASCII-128)
     * @var array
     */
    protected $alphabet;

    /**
     * case flag, whether upper case chars are preferred to lower case chars
     * @see self::CasePreferenceDefault
     * @var bool
     */
    protected $case;

    /**
     * @param array|string [optional] loose alphabet or key for predefined loose alphabet
     * @param string       [optional] encoding for multibyte functions
     * @param bool         [optional] prefer uppercase char to lower-case (Aa or aA)
     */
    public function __construct($encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault)
    {
        $this->setEncoding($encoding);
        $this->setUpperCasePreference($preferUpperCase);
        $this->loadAlphabet();
    }

    /**
     * @desc
     * The instance for all static calls is stored in a static property, so you can use the
     * instance once again. But if you prefer to use this class only once, use the __construct
     * method. There's no overhead then.
     */
    public static function getInstance($encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault)
    {
        $hash = crc32($encoding . $preferUpperCase);
        if (!array_key_exists($hash, self::$instances)) {
            self::$instances[$hash] = new self($encoding, $preferUpperCase);
        }
        return self::$instances[$hash];
    }

    /**
     * @desc shortcut
     */
    public static function sortStatic(&$array, $encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault)
    {
        $self = self::getInstance($encoding, $preferUpperCase);
        $self->sort($array);
    }

    public static function asortStatic(&$array, $encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault)
    {
        $self = self::getInstance($encoding, $preferUpperCase);
        $self->asort($array);
    }

    /**
     * @desc sort array with the constructor-given settings
     */
    public function sort(&$array)
    {
        /**
         * Got this message: "Warning: usort() [function.usort]: Invalid comparison function"?
         * Then do not call this method from a static context and change your error_reporting level
         * to E_ALL | E_STRICT.
         */
        usort($array, array($this, "sortString"));
    }

    public function asort(&$array)
    {
        uasort($array, array($this, "sortString"));
    }

    /**
     * @desc multibyte string compare (something like we would expect behind mb_strcmp())
     * @return int sort order value (1, 0, -1)
     */
    protected function sortString($a, $b)
    {
        $ax = mb_strlen($a, $this->encoding);
        $bx = mb_strlen($b, $this->encoding);
        for ($i = 0, $x = min($ax, $bx); $i < $x; ++$i) {
            $result = $this->charCmp(mb_substr($a, $i, 1, $this->encoding),
                                     mb_substr($b, $i, 1, $this->encoding));
            if ($result != 0) {
                return $result;
            }
        }
        return $this->intCmp($ax, $bx);
    }

    /**
     * @desc integer compare
     * @return int sort order value (-1, 0, 1)
     */
    protected function intCmp($a, $b)
    {
        return ($a == $b) ? 0 : ($a < $b ? -1 : 1);
    }

    /**
     * @desc multibyte char compare
     * @return int sort order value (-1, 0, 1)
     */
    protected function charCmp($a, $b)
    {
        // check if characters are known as special chars
        $ai = isset($this->alphabet[$a]); // ai = a isset
        $bi = isset($this->alphabet[$b]);

        if ($ai && $bi) {
            // both are known special chars
            $ar = $this->alphabet[$a]; // ar = a representation (ASCII-128)
            $br = $this->alphabet[$b];
            $result = $this->charCaseCmp($ar, $br);
            if ($result == 0 && $a != $b) {
                // they aren't equal, but their representation is, so check position in original array
                $ap = mb_strpos(self::$alphabetLoose[$ar], $a, 0, $this->encoding);
                $bp = mb_strpos(self::$alphabetLoose[$br], $b, 0, $this->encoding);
                $result = $this->intCmp($ap, $bp);
            }
        } else if ($ai) {
            // $a is a known special char, $b not
            $result = $this->charCaseCmp($this->alphabet[$a], $b);
             // so they are not equal, $result = 0 means $b is "smaller"
            $result = ($result == 0) ? 1 : $result;
        } else if ($bi) {
            // $b is a known special char, $a not
            $result = $this->charCaseCmp($a, $this->alphabet[$b]);
            // so they are not equal; $result = 0 means $a is "smaller"
            $result = ($result == 0) ? -1 : $result;
        } else {
            // both are unknown characters
            $result = $this->charCaseCmp($a, $b);
        }
        return $result;
    }

    /**
     * @desc multibyte char case compare
     * @return sort order value (-1, 0, 1)
     */
    protected function charCaseCmp($a, $b)
    {
        if ($a == $b) {
            // they are equal, no check required
            return 0;
        }
        $A = mb_strtoupper($a, $this->encoding);
        $B = mb_strtoupper($b, $this->encoding);
        $result = strcmp($A, $B);
        if ($result == 0) {
            // their mb_strtoupper() value is equal, select compare value depending on case preference
            $result = ($A != $a) ? $this->case : ($this->case * -1);
        }
        return $result;
    }

    protected function loadAlphabet()
    {
        if (empty($this->alphabet)) {
            // load only once per instance
            // $alphabetLoose is required to differ between special chars with the same ASCII-128 representation
            foreach (self::$alphabetLoose as $order => $char) {
                for ($i = 0, $x = mb_strlen($char, $this->encoding); $i < $x; ++$i) {
                    // use each multibyte char as key with its ASCII-128 representation as key
                    $this->alphabet[mb_substr($char, $i, 1, $this->encoding)] = $order;
                }
            }
        }
    }

    protected function setEncoding($encoding)
    {
        $this->encoding = $encoding;
    }

    protected function setUpperCasePreference($preferUpperCase)
    {
        $this->case = $preferUpperCase ? 1 : -1;
    }
}
?>

robo47 · 02.03.2009, 21:13

Also ich denke ein link zu phpclasses.org macht mehr Sinn, weil oben in der Zeile:

PHP-Code:

  protected static $alphabetLoose

hat es schon wieder was mit den Umlauten zerhauen

nikosch · 02.03.2009, 21:15

Ach ja? Bei mir nicht.

robo47 · 02.03.2009, 21:23

So siehts bei mir aus: Bild ansehen - LD-Host.de

so sollte es aussehen:
Bild ansehen - LD-Host.de

man beachte das "a" =>

Themen-Optionen
Druckbare Version zeigen Jemanden per E-Mail auf dieses Thema hinweisen
Thema bewerten
Thema bewerten:

Ähnliche Themen
Thema	Autor	Forum	Antworten	Letzter Beitrag
Turnierbaum aus array erstellen	kingflo	PHP-Fortgeschrittene	11	30.07.2008 11:32
Soapfault: Undefined Property	Argi	PHP-Fortgeschrittene	0	28.07.2008 11:17
ARRAY: alle möglichen Kombinationen	Tentacle	PHP Tipps 2006	13	12.05.2006 00:18
Übersicht aus mehrdimensionalem Array erzeugen	Crypi	PHP-Fortgeschrittene	0	30.01.2006 13:44
Array auslesen	andrew22	PHP Tipps 2006	3	21.01.2006 23:06
Array Formatierung		PHP Tipps 2006	3	17.01.2006 19:12
Menü mit Unterpunkten	supertramp	Beitragsarchiv	7	18.10.2005 22:40
Problem beim vergleichen von 2 Arrays		PHP Tipps 2005-2	1	06.10.2005 14:25
(schnellere) Funktion zum Zusammenfassen von CSS		PHP-Fortgeschrittene	21	08.08.2005 16:47
[Erledigt] Sortierung im Array mit Randomwerten		PHP-Fortgeschrittene	2	03.08.2005 09:28
Sortieren von Arrays mit mehr als 2 Dimensionen (Teil 2)	Buhmann	PHP-Fortgeschrittene	4	12.07.2005 14:03
[Erledigt] Mehrdimensionales Array in eindimensionales Array umwandeln		PHP-Fortgeschrittene	3	03.01.2005 22:31
Abfrage mit id aus anderer Tabelle	suter	PHP Tipps 2004-2	15	16.12.2004 14:25
Alphabetische Sortierung im Array		PHP Tipps 2004	4	17.08.2004 10:06

15.01.2009, 03:25
Chriz Moderator Registriert seit: 11.05.2008 Beiträge: 6.269	Multibyte Array Sortierung Hallo, ich habe ein Sortierskript für Multibyte-Arrays geschrieben, um endlich deutsche Umlaute unter UTF-8 gescheit sortiert zu bekommen. Ich würde mich freuen wenn ihr es mal testen oder mir Performance-Verbesserungen vorschlagen würdet. Gern auch Hinweise auf interne PHP-Funktionen die schon genau das können (sort und strcoll sortieren scheinbar noch strikt nach Zeichentabelleneintrag). Geändert von Chriz (15.01.2009 um 03:31 Uhr).


Mister Ad PHP Code Flüsterer Registriert seit: 21.08.2005 Beiträge: 4682 PHP-Kenntnisse: Fortgeschritten

15.01.2009, 03:27
Chriz Moderator Registriert seit: 11.05.2008 Beiträge: 6.269	Ja ich würds gern posten, aber ein Bug im Forum schneidet alles bei den UTF-8 Zeichen ab. Hier die base64 Version: PHP-Code: <?php echo base64_decode(<<<B64 <?php
error_reporting(E_ALL | E_STRICT);
header("Content-type: text/html; charset=utf-8");

/**
 * @desc
 * Sort arrays with multibyte values using a predefined or custom alphabet.
 *
 * @author Christian Reinecke <reinecke@bajoodoo.de>
 * @version 1.0
 * @since 2008-01-15
 *
 * @example
 * <code>
 *     // quick
 *     $array = array("Zebra", "Engländer", "Englander", "Chinese", "Äffchen", "Überseeboot",
 *                    "Unterseeboot", "ß", "Sz", "Bürger", "Burger", "burger", "bürger", "ÄÖÜ"));
 *     AIS_Util_MultibyteSort::staticSort($array);
 *     // Äffchen, ÄÖÜ, Burger, Bürger, burger, bürger, Chinese, Englander, Engländer, Sz, ß,
 *     // Unterseeboot, Überseeboot, Zebra
 *
 *
 *     // with custom setup
 *     $mbsort = new AIS_Util_MultibyteSort(AIS_Util_MultibyteSort::AlphabetGerman,
 *                                          AIS_Util_MultibyteSort::EncodingUtf8,
 *                                          AIS_Util_MultibyteSort::CasePreferenceUpper);
 *     $mbsort->sort($array);
 *     // same result as above
 *     $mbsort->sort($otherArray);
 * </code>
 *
 * @todo
 * This class can not handle unknown special characters (fallback to default behaviour),
 * which makes it desirable to find a global, sorted alphabet. Feel free to add more char maps.
 */
class AIS_Util_MultibyteSort
{
    const EncodingUtf8    = "UTF-8";
    const EncodingDefault = self::EncodingUtf8;

    const CasePreferenceUpper   = true;
    const CasePreferenceLower   = false;
    const CasePreferenceDefault = self::CasePreferenceUpper;

    const AlphabetGerman  = "german";
    const AlphabetDefault = self::AlphabetGerman;

    /**
     * predefined loose alphabets
     */
    protected static $alphabets = array(
        self::AlphabetGerman => array("A" => "Ä", "a" => "äà", "O" => "Ö", "o" => "ö", "U" => "Ü",
                                      "u" => "ü", "s" => "ß"));

    /**
     * multibyte encoding
     * @see self::EncodingDefault
     * @var string
     */
    protected $encoding;

    /**
     * used loose alphabet table
     * @see self::$alphabets
     * @var array
     */
    protected $alphabetLoose;

    /**
     * array map (single special char => ASCII-128)
     * @var array
     */
    protected $alphabet;

    /**
     * case flag, whether upper case chars are preferred to lower case chars
     * @see self::CasePreferenceDefault
     * @var bool
     */
    protected $case;

    /**
     * @param array|string [optional] loose alphabet or key for predefined loose alphabet
     * @param string       [optional] encoding for multibyte functions
     * @param bool         [optional] prefer uppercase char to lower-case (Aa or aA)
     */
    public function __construct($alphabet = self::AlphabetDefault, $encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault)
    {
        $this->setEncoding($encoding);
        $this->setAlphabet($alphabet);
        $this->setUpperCasePreference($preferUpperCase);
    }

    /**
     * @desc shortcut
     */
    public static function staticSort(&$array, $alphabet = self::AlphabetDefault, $encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault)
    {
        $self = new self($alphabet, $encoding, $preferUpperCase);
        $self->sort($array);
    }

    /**
     * @desc sort array with the constructor-given settings
     */
    public function sort(&$array)
    {
        /**
         * Got this message: "Warning: usort() [function.usort]: Invalid comparison function"?
         * Then do not call this method from a static context or change your error_reporting level
         * to E_ALL | E_STRICT.
         */
        usort($array, array($this, "sortString"));
    }

    /**
     * @desc multibyte string compare (something like we would expect behind mb_strcmp())
     * @return int sort order value (1, 0, -1)
     */
    protected function sortString($a, $b)
    {
        $ax = mb_strlen($a, $this->encoding);
        $bx = mb_strlen($b, $this->encoding);
        for ($i = 0, $x = min($ax, $bx); $i < $x; ++$i) {
            $result = $this->charCmp(mb_substr($a, $i, 1, $this->encoding),
                                     mb_substr($b, $i, 1, $this->encoding));
            if ($result != 0) {
                return $result;
            }
        }
        return $this->intCmp($ax, $bx);
    }

    /**
     * @desc integer compare
     * @return int sort order value (-1, 0, 1)
     */
    protected function intCmp($a, $b)
    {
        return ($a == $b) ? 0 : ($a < $b ? -1 : 1);
    }

    /**
     * @desc multibyte char compare
     * @return int sort order value (-1, 0, 1)
     */
    protected function charCmp($a, $b)
    {
        // check if characters are known as special chars
        $ai = isset($this->alphabet[$a]); // ai = a isset
        $bi = isset($this->alphabet[$b]);

        if ($ai && $bi) {
            // both are known special chars
            $ar = $this->alphabet[$a]; // ar = a representation (ASCII-128)
            $br = $this->alphabet[$b];
            $result = $this->charCaseCmp($ar, $br);
            if ($result == 0 && $a != $b) {
                // they aren't equal, but their representation is, so check position in original array
                $ap = mb_strpos($this->alphabetLoose[$ar], $a, 0, $this->encoding);
                $bp = mb_strpos($this->alphabetLoose[$br], $b, 0, $this->encoding);
                $result = $this->intCmp($ap, $bp);
            }
        } else if ($ai) {
            // $a is a known special char, $b not
            $result = $this->charCaseCmp($this->alphabet[$a], $b);
             // so they are not equal, $result = 0 means $b is "smaller"
            $result = ($result == 0) ? 1 : $result;
        } else if ($bi) {
            // $b is a known special char, $a not
            $result = $this->charCaseCmp($a, $this->alphabet[$b]);
            // so they are not equal; $result = 0 means $a is "smaller"
            $result = ($result == 0) ? -1 : $result;
        } else {
            // both are unknown characters
            $result = $this->charCaseCmp($a, $b);
        }
        return $result;
    }

    /**
     * @desc multibyte char case compare
     * @return sort order value (-1, 0, 1)
     */
    protected function charCaseCmp($a, $b)
    {
        if ($a == $b) {
            // they are equal, no check required
            return 0;
        }
        $A = mb_strtoupper($a, $this->encoding);
        $B = mb_strtoupper($b, $this->encoding);
        $result = strcmp($A, $B);
        if ($result == 0) {
            // their mb_strtoupper() value is equal, select compare value depending on case preference
            $result = ($A != $a) ? $this->case : ($this->case * -1);
        }
        return $result;
    }

    protected function setAlphabet($alphabet)
    {
        if (!is_array($alphabet)) {
            // use $alphabet as key for predefined alphabets
            $alphabet = self::$alphabets[$alphabet];
        }
        // store loose alphabet to differ between special chars with the same ASCII-128 representation
        $this->alphabetLoose = $alphabet;
        foreach ($this->alphabetLoose as $order => $char) {
            for ($i = 0, $x = mb_strlen($char, $this->encoding); $i < $x; ++$i) {
                // use each multibyte char as key with its ASCII-128 representation as key
                $this->alphabet[mb_substr($char, $i, 1, $this->encoding)] = $order;
            }
        }
    }

    protected function setEncoding($encoding)
    {
        $this->encoding = $encoding;
    }

    protected function setUpperCasePreference($preferUpperCase)
    {
        $this->case = $preferUpperCase ? 1 : -1;
    }
} B64 ); ?> Geändert von Chriz (15.01.2009 um 03:33 Uhr).

15.01.2009, 12:27
nikosch moderatives Dielektrikum Registriert seit: 21.05.2008 Beiträge: 35.992 PHP-Kenntnisse: Fortgeschritten	Cool. Ein Geheimposting! PS: Schöne Sache. Ich würde noch case-insensitivity anregen. __________________ -- One pixel is still too big. Please make it smaller. ASAP. Initiative Mittelstand. Die wichtigste Gestaltungsregel im Screendesign ist Pi mal Daumen des Arbeitgebers. -- Geändert von nikosch (15.01.2009 um 12:33 Uhr).

15.01.2009, 13:54
Chriz Moderator Registriert seit: 11.05.2008 Beiträge: 6.269	PHP-Code: <?php /** * @desc * Sort arrays with multibyte values using a predefined or custom alphabet. * * @author Christian Reinecke <reinecke@bajoodoo.de> * @version 1.0 * @since 2008-01-15 * * @todo * This class can not handle unknown special characters (fallback to default behaviour), * which makes it desirable to find a global, sorted alphabet. Feel free to add more char maps. / class AIS_Util_MultibyteSort { const EncodingUtf8 = "UTF-8"; const EncodingDefault = self::EncodingUtf8; const CasePreferenceUpper = true; const CasePreferenceLower = false; const CasePreferenceDefault = self::CasePreferenceUpper; /* * multibyte encoding * @see self::EncodingDefault * @var string / protected $encoding; /* * used loose alphabet table * @see self::$alphabets * @var array / protected $alphabetLoose; /* * array map (single special char => ASCII-128) * @var array / protected $alphabet; /* * case flag, whether upper case chars are preferred to lower case chars * @see self::CasePreferenceDefault * @var bool / protected $case; /* * @param array\|string [optional] loose alphabet or key for predefined loose alphabet * @param string [optional] encoding for multibyte functions * @param bool [optional] prefer uppercase char to lower-case (Aa or aA) / public function __construct($alphabet, $encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault) { $this->setEncoding($encoding); $this->setAlphabet($alphabet); $this->setUpperCasePreference($preferUpperCase); } /* * @desc shortcut / public static function staticSort(&$array, $alphabet, $encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault) { $self = new self($alphabet, $encoding, $preferUpperCase); $self->sort($array); } /* * @desc sort array with the constructor-given settings / public function sort(&$array) { /* * Got this message: "Warning: usort() [function.usort]: Invalid comparison function"? * Then do not call this method from a static context or change your error_reporting level * to E_ALL \| E_STRICT. / usort($array, array($this, "sortString")); } /* * @desc multibyte string compare (something like we would expect behind mb_strcmp()) * @return int sort order value (1, 0, -1) / protected function sortString($a, $b) { $ax = mb_strlen($a, $this->encoding); $bx = mb_strlen($b, $this->encoding); for ($i = 0, $x = min($ax, $bx); $i < $x; ++$i) { $result = $this->charCmp(mb_substr($a, $i, 1, $this->encoding), mb_substr($b, $i, 1, $this->encoding)); if ($result != 0) { return $result; } } return $this->intCmp($ax, $bx); } /* * @desc integer compare * @return int sort order value (-1, 0, 1) / protected function intCmp($a, $b) { return ($a == $b) ? 0 : ($a < $b ? -1 : 1); } /* * @desc multibyte char compare * @return int sort order value (-1, 0, 1) / protected function charCmp($a, $b) { // check if characters are known as special chars $ai = isset($this->alphabet[$a]); // ai = a isset $bi = isset($this->alphabet[$b]); if ($ai && $bi) { // both are known special chars $ar = $this->alphabet[$a]; // ar = a representation (ASCII-128) $br = $this->alphabet[$b]; $result = $this->charCaseCmp($ar, $br); if ($result == 0 && $a != $b) { // they aren't equal, but their representation is, so check position in original array $ap = mb_strpos($this->alphabetLoose[$ar], $a, 0, $this->encoding); $bp = mb_strpos($this->alphabetLoose[$br], $b, 0, $this->encoding); $result = $this->intCmp($ap, $bp); } } else if ($ai) { // $a is a known special char, $b not $result = $this->charCaseCmp($this->alphabet[$a], $b); // so they are not equal, $result = 0 means $b is "smaller" $result = ($result == 0) ? 1 : $result; } else if ($bi) { // $b is a known special char, $a not $result = $this->charCaseCmp($a, $this->alphabet[$b]); // so they are not equal; $result = 0 means $a is "smaller" $result = ($result == 0) ? -1 : $result; } else { // both are unknown characters $result = $this->charCaseCmp($a, $b); } return $result; } /* * @desc multibyte char case compare * @return sort order value (-1, 0, 1) / protected function charCaseCmp($a, $b) { if ($a == $b) { // they are equal, no check required return 0; } $A = mb_strtoupper($a, $this->encoding); $B = mb_strtoupper($b, $this->encoding); $result = strcmp($A, $B); if ($result == 0) { // their mb_strtoupper() value is equal, select compare value depending on case preference $result = ($A != $a) ? $this->case : ($this->case -1); } return $result; } protected function setAlphabet($alphabet) { if (!is_array($alphabet)) { // use $alphabet as key for predefined alphabets $alphabet = self::$alphabets[$alphabet]; } // store loose alphabet to differ between special chars with the same ASCII-128 representation $this->alphabetLoose = $alphabet; foreach ($this->alphabetLoose as $order => $char) { for ($i = 0, $x = mb_strlen($char, $this->encoding); $i < $x; ++$i) { // use each multibyte char as key with its ASCII-128 representation as key $this->alphabet[mb_substr($char, $i, 1, $this->encoding)] = $order; } } } protected function setEncoding($encoding) { $this->encoding = $encoding; } protected function setUpperCasePreference($preferUpperCase) { $this->case = $preferUpperCase ? 1 : -1; } } ?> Geändert von Chriz (15.01.2009 um 14:01 Uhr).

15.01.2009, 14:04
Chriz Moderator Registriert seit: 11.05.2008 Beiträge: 6.269	Gibts doch nicht, ich kann das Anwendungsbeispiel nichtmal eingetippt hier posten OK dann zur Erklärung, ein $alphabet ist ein assoziativer Array, der Schlüssel muss ein ASCII-128-Zeichen sein, dessen Werte die Sonderzeichen, die bei der Sortierung dann direkt nach diesem ASCII-128-Zeichen einsortiert werden. Möglich ist als Wert mehr als ein Zeichen, also ä oder Akcent a (das kann ich hier offenbar nicht posten). Geändert von Chriz (15.01.2009 um 14:07 Uhr).

15.01.2009, 14:22
nikosch moderatives Dielektrikum Registriert seit: 21.05.2008 Beiträge: 35.992 PHP-Kenntnisse: Fortgeschritten	Yepp, bekanntes Problem: http://www.php.de/board-support/4935...uckt-text.html (Zeichensatz-Probleme, Board verschluckt Text) __________________ -- One pixel is still too big. Please make it smaller. ASAP. Initiative Mittelstand. Die wichtigste Gestaltungsregel im Screendesign ist Pi mal Daumen des Arbeitgebers. --

02.03.2009, 20:13
Chriz Moderator Registriert seit: 11.05.2008 Beiträge: 6.269	Hier nochmal die verständliche Version mit erweitertem Kommentar und Beispiel, ich hoffe das klappt jetzt - habe das Akcent a entfernt. Die API hat sich etwas verändert, das Alphabet muss jetzt nicht mehr übergeben werden, sondern ist vordefiniert. PHP-Code: <?php /** * @desc * Sort arrays with multibyte values using a predefined or custom alphabet. * * @author Christian Reinecke <reinecke@bajoodoo.de> * @version 1.1 * @since 2009-01-15 * * @example * <code> * // quick * $array = array("Zebra", "Engländer", "England", "Chinese", "Äffchen", "Überseeboot", * "Unterseeboot", "ß", "Sz", "Bürger", "Burger", "burger", "bürger", "ÄÖÜ"); * AIS_Util_MultibyteSort::staticSort($array); * print_r($array); * // Array ( * // [0] => Äffchen [1] => ÄÖÜ [2] => Burger [3] => Bürger [4] => burger [5] => bürger * // [6] => Chinese [7] => England [8] => Engländer [9] => Sz [10] => ß [11] => Unterseeboot * // [12] => Überseeboot [13] => Zebra * // ) * * * // with custom setup * $mbsort = new AIS_Util_MultibyteSort("UTF-8", true); * $mbsort->sort($array); * print_r($array) * // same result as above * $mbsort->sort($otherArray); * </code> * * @todo * This class can not handle unknown special characters (fallback to default behaviour), * which makes it desirable to find a global, sorted alphabet. Feel free to add more special chars. * * Don't forget to send the correct charset in your header * <code> * header("Content-type: text/html; charset=utf-8"); * </code> / class AIS_Util_MultibyteSort { const EncodingUtf8 = "UTF-8"; const EncodingDefault = self::EncodingUtf8; const CasePreferenceUpper = true; const CasePreferenceLower = false; const CasePreferenceDefault = self::CasePreferenceUpper; protected static $instances = array(); protected static $alphabetLoose = array("A" => "Ä", "a" => "äæå", "O" => "Ö", "o" => "ö", "U" => "Ü", "u" => "ü", "s" => "ß"); /* * multibyte encoding * @see self::EncodingDefault * @var string / protected $encoding; /* * array map (single special char => ASCII-128) * @var array / protected $alphabet; /* * case flag, whether upper case chars are preferred to lower case chars * @see self::CasePreferenceDefault * @var bool / protected $case; /* * @param array\|string [optional] loose alphabet or key for predefined loose alphabet * @param string [optional] encoding for multibyte functions * @param bool [optional] prefer uppercase char to lower-case (Aa or aA) / public function __construct($encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault) { $this->setEncoding($encoding); $this->setUpperCasePreference($preferUpperCase); $this->loadAlphabet(); } /* * @desc * The instance for all static calls is stored in a static property, so you can use the * instance once again. But if you prefer to use this class only once, use the __construct * method. There's no overhead then. / public static function getInstance($encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault) { $hash = crc32($encoding . $preferUpperCase); if (!array_key_exists($hash, self::$instances)) { self::$instances[$hash] = new self($encoding, $preferUpperCase); } return self::$instances[$hash]; } /* * @desc shortcut / public static function sortStatic(&$array, $encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault) { $self = self::getInstance($encoding, $preferUpperCase); $self->sort($array); } public static function asortStatic(&$array, $encoding = self::EncodingDefault, $preferUpperCase = self::CasePreferenceDefault) { $self = self::getInstance($encoding, $preferUpperCase); $self->asort($array); } /* * @desc sort array with the constructor-given settings / public function sort(&$array) { /* * Got this message: "Warning: usort() [function.usort]: Invalid comparison function"? * Then do not call this method from a static context and change your error_reporting level * to E_ALL \| E_STRICT. / usort($array, array($this, "sortString")); } public function asort(&$array) { uasort($array, array($this, "sortString")); } /* * @desc multibyte string compare (something like we would expect behind mb_strcmp()) * @return int sort order value (1, 0, -1) / protected function sortString($a, $b) { $ax = mb_strlen($a, $this->encoding); $bx = mb_strlen($b, $this->encoding); for ($i = 0, $x = min($ax, $bx); $i < $x; ++$i) { $result = $this->charCmp(mb_substr($a, $i, 1, $this->encoding), mb_substr($b, $i, 1, $this->encoding)); if ($result != 0) { return $result; } } return $this->intCmp($ax, $bx); } /* * @desc integer compare * @return int sort order value (-1, 0, 1) / protected function intCmp($a, $b) { return ($a == $b) ? 0 : ($a < $b ? -1 : 1); } /* * @desc multibyte char compare * @return int sort order value (-1, 0, 1) / protected function charCmp($a, $b) { // check if characters are known as special chars $ai = isset($this->alphabet[$a]); // ai = a isset $bi = isset($this->alphabet[$b]); if ($ai && $bi) { // both are known special chars $ar = $this->alphabet[$a]; // ar = a representation (ASCII-128) $br = $this->alphabet[$b]; $result = $this->charCaseCmp($ar, $br); if ($result == 0 && $a != $b) { // they aren't equal, but their representation is, so check position in original array $ap = mb_strpos(self::$alphabetLoose[$ar], $a, 0, $this->encoding); $bp = mb_strpos(self::$alphabetLoose[$br], $b, 0, $this->encoding); $result = $this->intCmp($ap, $bp); } } else if ($ai) { // $a is a known special char, $b not $result = $this->charCaseCmp($this->alphabet[$a], $b); // so they are not equal, $result = 0 means $b is "smaller" $result = ($result == 0) ? 1 : $result; } else if ($bi) { // $b is a known special char, $a not $result = $this->charCaseCmp($a, $this->alphabet[$b]); // so they are not equal; $result = 0 means $a is "smaller" $result = ($result == 0) ? -1 : $result; } else { // both are unknown characters $result = $this->charCaseCmp($a, $b); } return $result; } /* * @desc multibyte char case compare * @return sort order value (-1, 0, 1) / protected function charCaseCmp($a, $b) { if ($a == $b) { // they are equal, no check required return 0; } $A = mb_strtoupper($a, $this->encoding); $B = mb_strtoupper($b, $this->encoding); $result = strcmp($A, $B); if ($result == 0) { // their mb_strtoupper() value is equal, select compare value depending on case preference $result = ($A != $a) ? $this->case : ($this->case -1); } return $result; } protected function loadAlphabet() { if (empty($this->alphabet)) { // load only once per instance // $alphabetLoose is required to differ between special chars with the same ASCII-128 representation foreach (self::$alphabetLoose as $order => $char) { for ($i = 0, $x = mb_strlen($char, $this->encoding); $i < $x; ++$i) { // use each multibyte char as key with its ASCII-128 representation as key $this->alphabet[mb_substr($char, $i, 1, $this->encoding)] = $order; } } } } protected function setEncoding($encoding) { $this->encoding = $encoding; } protected function setUpperCasePreference($preferUpperCase) { $this->case = $preferUpperCase ? 1 : -1; } } ?>

02.03.2009, 21:13
robo47 Moderator Registriert seit: 03.09.2004 Beiträge: 11.792 PHP-Kenntnisse: Fortgeschritten	Also ich denke ein link zu phpclasses.org macht mehr Sinn, weil oben in der Zeile: PHP-Code: `protected static $alphabetLoose` hat es schon wieder was mit den Umlauten zerhauen __________________ robo47.net - Blog, Codeschnipsel und mehr \| Caching-Klassen und Opcode Caches in php \| Robo47 Components - PHP Library extending Zend Framework

02.03.2009, 21:15
nikosch moderatives Dielektrikum Registriert seit: 21.05.2008 Beiträge: 35.992 PHP-Kenntnisse: Fortgeschritten	Ach ja? Bei mir nicht. __________________ -- One pixel is still too big. Please make it smaller. ASAP. Initiative Mittelstand. Die wichtigste Gestaltungsregel im Screendesign ist Pi mal Daumen des Arbeitgebers. --

02.03.2009, 21:23
robo47 Moderator Registriert seit: 03.09.2004 Beiträge: 11.792 PHP-Kenntnisse: Fortgeschritten	So siehts bei mir aus: Bild ansehen - LD-Host.de so sollte es aussehen: Bild ansehen - LD-Host.de man beachte das "a" => __________________ robo47.net - Blog, Codeschnipsel und mehr \| Caching-Klassen und Opcode Caches in php \| Robo47 Components - PHP Library extending Zend Framework