could anyone make this function run more faster?

[itmanvn]

function VN2none($str)
{
	//khoi tao string
	$vstr = 'aAeEoOuUiIdDyY';
	$Thga = 'áàạảãâấầậẩẫăắằặẳẵ';
	$HoaA = 'ÁÀẠẢÃÂẤẦẬẨẪĂẮẰẶẲẴ';
	$Thge = 'éèẹẻẽêếềệểễeeeeee';
	$HoaE = 'ÉÈẸẺẼÊẾỀỆỂỄEEEEEE';
	$Thgo = 'óòọỏõôốồộổỗơớờợởỡ';
	$HoaO = 'ÓÒỌỎÕÔỐỒỘỔỖƠỚỜỢỞỠ';
	$Thgu = 'úùụủũưứừựửữuuuuuu';
	$HoaU = 'ÚÙỤỦŨƯỨỪỰỬỮUUUUUU';
	$Thgi = 'íìịỉĩiiiiiiiiiiii';
	$HoaI = 'ÍÌỊỈĨIIIIIIIIIIII';
	$Thgd = 'đdddddddddddddddd';
	$HoaD = 'ĐDDDDDDDDDDDDDDDD';
	$Thgy = 'ýỳỵỷỹyyyyyyyyyyyy';
	$HoaY = 'ÝỲỴỶỸYYYYYYYYYYYY';
	
	// khoi tao mang
	for ($i = 0; $i <= 13; $i++)
	{
		$arr_matrix[$i][0] = utf8_substr($vstr, $i, 1);
	}
	
	// nap ma tran
	for ($j = 1; $j <= 17; $j++)
	{
		for ($i = 1; $i <= 17; $i++)
		{
			$arr_matrix[0][$i] = utf8_substr($Thga, $i-1, 1);
			$arr_matrix[1][$i] = utf8_substr($HoaA, $i-1, 1);
			$arr_matrix[2][$i] = utf8_substr($Thge, $i-1, 1);
			$arr_matrix[3][$i] = utf8_substr($HoaE, $i-1, 1);
			$arr_matrix[4][$i] = utf8_substr($Thgo, $i-1, 1);
			$arr_matrix[5][$i] = utf8_substr($HoaO, $i-1, 1);
			$arr_matrix[6][$i] = utf8_substr($Thgu, $i-1, 1);
			$arr_matrix[7][$i] = utf8_substr($HoaU, $i-1, 1);
			$arr_matrix[8][$i] = utf8_substr($Thgi, $i-1, 1);
			$arr_matrix[9][$i] = utf8_substr($HoaI, $i-1, 1);
			$arr_matrix[10][$i] = utf8_substr($Thgd, $i-1, 1);
			$arr_matrix[11][$i] = utf8_substr($HoaD, $i-1, 1);
			$arr_matrix[12][$i] = utf8_substr($Thgy, $i-1, 1);
			$arr_matrix[13][$i] = utf8_substr($HoaY, $i-1, 1);
		}
	}
	
	// loai bo dau
	$str = trim($str);
	for ($j = 0; $j <= 13; $j++)
	{
		for ($i = 1; $i <= 17; $i++)
		{
			$str = str_replace($arr_matrix[$j][$i], $arr_matrix[$j][0], $str);
		}
	}
	$str = url_title($str, '_');
	return $str;
}

function utf8_substr($str, $from, $len)
{
	return preg_replace('#^(?:[\x00-\x7F]|[\xC0-\xFF][\x80-\xBF]+){0,'.$from.'}'.
                       '((?:[\x00-\x7F]|[\xC0-\xFF][\x80-\xBF]+){0,'.$len.'}).*#s',
                       '$1',$str);
}

[Mittineague]

The only thing I see that might make it (IMHO insignificantly) faster would be to use a variable to hold the $i-1 value instead of calculating it each line. ie.

        for ($i = 1; $i <= 17; $i++)
        {
            $value = $i-1;
            $arr_matrix[0][$i] = utf8_substr($Thga, $value, 1);
            $arr_matrix[1][$i] = utf8_substr($HoaA, $value, 1);
            $arr_matrix[2][$i] = utf8_substr($Thge, $value, 1);
...... 

I imagine the regex is the main consumer, not much you can do about that.

[kromey]

Looking at your code, $arr_matrix is going to be identical each and every time this function is run. So probably the single best thing you could do to improve the speed is to not regenerate it each time but to simply hard-code it into your function. If you don't like that, try adding some basic memoization to your function, although this will only improve your speed if it is called multiple times per script.

I really think just hard-coding $arr_matrix instead of regenerating the whole thing each time is by far your best bet.

[bokehman]

Here's one way.

function VN2none($str)
{
    //khoi tao string
    static $vstr = 'aAeEoOuUiIdDyY';
    static $Thga = 'áà??ãâ?????a?????';
    static $HoaA = 'ÁÀ??ÃÂ?????A?????';
    static $Thge = 'éè???ê?????eeeeee';
    static $HoaE = 'ÉÈ???Ê?????EEEEEE';
    static $Thgo = 'óò??õô?????o?????';
    static $HoaO = 'ÓÒ??ÕÔ?????O?????';
    static $Thgu = 'úù??uu?????uuuuuu';
    static $HoaU = 'ÚÙ??UU?????UUUUUU';
    static $Thgi = 'íì??iiiiiiiiiiiii';
    static $HoaI = 'ÍÌ??IIIIIIIIIIIII';
    static $Thgd = 'ddddddddddddddddd';
    static $HoaD = 'ÐDDDDDDDDDDDDDDDD';
    static $Thgy = 'ý????yyyyyyyyyyyy';
    static $HoaY = 'Ý????YYYYYYYYYYYY';
    static $arr_matrix = array();
    
    if(!$arr_matrix)
    {
        // khoi tao mang
        for ($i = 0; $i <= 13; $i++)
        {
            $arr_matrix[$i][0] = utf8_substr($vstr, $i, 1);
        }
        
        // nap ma tran
        for ($j = 1; $j <= 17; $j++)
        {
            for ($i = 1; $i <= 17; $i++)
            {
                $arr_matrix[0][$i] = utf8_substr($Thga, $i-1, 1);
                $arr_matrix[1][$i] = utf8_substr($HoaA, $i-1, 1);
                $arr_matrix[2][$i] = utf8_substr($Thge, $i-1, 1);
                $arr_matrix[3][$i] = utf8_substr($HoaE, $i-1, 1);
                $arr_matrix[4][$i] = utf8_substr($Thgo, $i-1, 1);
                $arr_matrix[5][$i] = utf8_substr($HoaO, $i-1, 1);
                $arr_matrix[6][$i] = utf8_substr($Thgu, $i-1, 1);
                $arr_matrix[7][$i] = utf8_substr($HoaU, $i-1, 1);
                $arr_matrix[8][$i] = utf8_substr($Thgi, $i-1, 1);
                $arr_matrix[9][$i] = utf8_substr($HoaI, $i-1, 1);
                $arr_matrix[10][$i] = utf8_substr($Thgd, $i-1, 1);
                $arr_matrix[11][$i] = utf8_substr($HoaD, $i-1, 1);
                $arr_matrix[12][$i] = utf8_substr($Thgy, $i-1, 1);
                $arr_matrix[13][$i] = utf8_substr($HoaY, $i-1, 1);
            }
        }
    }
    
    // loai bo dau
    $str = trim($str);
    for ($j = 0; $j <= 13; $j++)
    {
        for ($i = 1; $i <= 17; $i++)
        {
            $str = str_replace($arr_matrix[$j][$i], $arr_matrix[$j][0], $str);
        }
    }
    $str = url_title($str, '_');
    return $str;
} 

Looking at your code, $arr_matrix is going to be identical each and every time this function is run. So probably the single best thing you could do to improve the speed is to not regenerate it each time but to simply hard-code it into your function.

I agree but if you go down that road make it a static variable too otherwise it still has to be built each time the function rolls which in a big loop really uses a lot of effort.

[itmanvn]

thank you all for replying me, after ask on some forum I got this solution

function VN2none ($str) {
static $from =
          'áàạảãâấầậẩẫăắằặẳẵ'.
          'ÁÀẠẢÃÂẤẦẬẨẪĂẮẰẶẲẴ'.
          'éèẹẻẽêếềệểễeeeeee'.
          'ÉÈẸẺẼÊẾỀỆỂỄEEEEEE'.
          'óòọỏõôốồộổỗơớờợởỡ'.
          'ÓÒỌỎÕÔỐỒỘỔỖƠỚỜỢỞỠ'.
          'úùụủũưứừựửữuuuuuu'.
          'ÚÙỤỦŨƯỨỪỰỬỮUUUUUU'.
          'íìịỉĩiiiiiiiiiiii'.
          'ÍÌỊỈĨIIIIIIIIIIII'.
          'đdddddddddddddddd'.
          'ĐDDDDDDDDDDDDDDDD'.
          'ýỳỵỷỹyyyyyyyyyyyy'.
          'ÝỲỴỶỸYYYYYYYYYYYY',
       $to =
          'aaaaaaaaaaaaaaaaa'.
          'AAAAAAAAAAAAAAAAA'.
          'eeeeeeeeeeeeeeeee'.
          'EEEEEEEEEEEEEEEEE'.
          'ooooooooooooooooo'.
          'OOOOOOOOOOOOOOOOO'.
          'uuuuuuuuuuuuuuuuu'.
          'UUUUUUUUUUUUUUUUU'.
          'iiiiiiiiiiiiiiiii'.
          'IIIIIIIIIIIIIIIII'.
          'ddddddddddddddddd'.
          'DDDDDDDDDDDDDDDDD'.
          'yyyyyyyyyyyyyyyyy'.
          'YYYYYYYYYYYYYYYYY';

    return url_title(strtr(trim($str), $from, $to), '_');
} 

note that: this will convert all vietnamese character to an english character in a string. Example: “chào các bạn” --> “chao cac ban”.

[bokehman]

thank you all for replying me, after ask on some forum I got this solution

function VN2none ($str) {
static $from =
          'áàạảãâấầậẩẫăắằặẳẵ'.
          'ÁÀẠẢÃÂẤẦẬẨẪĂẮẰẶẲẴ'.
          'éèẹẻẽêếềệểễeeeeee'.
          'ÉÈẸẺẼÊẾỀỆỂỄEEEEEE'.
          'óòọỏõôốồộổỗơớờợởỡ'.
          'ÓÒỌỎÕÔỐỒỘỔỖƠỚỜỢỞỠ'.
          'úùụủũưứừựửữuuuuuu'.
          'ÚÙỤỦŨƯỨỪỰỬỮUUUUUU'.
          'íìịỉĩiiiiiiiiiiii'.
          'ÍÌỊỈĨIIIIIIIIIIII'.
          'đdddddddddddddddd'.
          'ĐDDDDDDDDDDDDDDDD'.
          'ýỳỵỷỹyyyyyyyyyyyy'.
          'ÝỲỴỶỸYYYYYYYYYYYY',
       $to =
          'aaaaaaaaaaaaaaaaa'.
          'AAAAAAAAAAAAAAAAA'.
          'eeeeeeeeeeeeeeeee'.
          'EEEEEEEEEEEEEEEEE'.
          'ooooooooooooooooo'.
          'OOOOOOOOOOOOOOOOO'.
          'uuuuuuuuuuuuuuuuu'.
          'UUUUUUUUUUUUUUUUU'.
          'iiiiiiiiiiiiiiiii'.
          'IIIIIIIIIIIIIIIII'.
          'ddddddddddddddddd'.
          'DDDDDDDDDDDDDDDDD'.
          'yyyyyyyyyyyyyyyyy'.
          'YYYYYYYYYYYYYYYYY';

    return url_title(strtr(trim($str), $from, $to), '_');
} 

note that: this will convert all vietnamese character to an english character in a string. Example: “chào các bạn” --> “chao cac ban”.

Make your $to variable static also.

[kromey]

Just out of curiosity, how much faster does this algorithm run now? I've no doubt it's significant, I'm just interested in knowing by how much.