ereg_replace problem

[Gerlof]

Hi i got this code:

        function removedata2($str, $a=false)
        {
        $str = ereg_replace("[^A-Za-z0-9]", "-", $str);
        $str = str_replace("'","", $str);
        $str = str_replace("&","&", $str);
        $str = strip_tags($str, '');
        if ($a==true && strlen($str)>=1000000){
        $str=substr($str, 0, 1000000)."..";
        }
        return $str;
        } 

It replaces ahref url to something like hello-my-name-is.
But my code got weird bug with ereg_replace.
That it makes: hello---my-name---is-
So sometimes it does -- and sometimes one -
Looks like he replaces every char expect AZ az 09

Is there way to do this nicer? And cleaner.

So i get nice url title with single - in it?

[spikeZ]

Isnt this line:

// removed carat ^ from regex
$str = ereg_replace("[A-Za-z0-9]", "-", $str); 

going to result in:

-----------------

if it replaces all the Alphanumeric characters with a - ?

What are you trying to do Gerlof?

[kleineme]

If you have multiple non-alphanumeric characters after each other, then you will get multiple hyphens instead. If you think that ereg_replace is behaving irregularly then please post some examples of $str with the corresponding return values of removedata2.

By the way, the two calls to str_replace and the one to strip_tags are superfluous because after ereg_replace there will be no more special characters in your string which could be replaced or stripped.

edit: I only see this now:

Looks like he replaces every char expect AZ az 09

Exactly, that's what you told ereg_replace to do

[logic_earth]

function removedata2($str, $a = false) {
 
    $str = preg_replace(array('/[^a-z0-9]+/i', '/-{2,}/'), '-', $str);
 
    if ($a && strlen($str) >= 1000000) {
        $str = substr($str, 0, 1000000) . '..';
    }
 
    return $str;
}

preg_* uses Perl-compatible syntax using the PCRE functions which is more widely supported then POSIX Extended at least of what I seen.

Edit:
It has come to me attention that the regex function can be simplified down to:

function removedata2($str, $a = false) {
 
    $str = preg_replace('/[^a-z0-9]+/i', '-', $str);
 
    if ($a && strlen($str) >= 1000000) {
        $str = substr($str, 0, 1000000) . '..';
    }
 
    return $str;
}

No need for the second regex '/-{2,}/'