Don't use regular expressions when dealing with URL values. It's much easier (and safer) to handle them as a URL instead of plain text.
This could be one way to do it:
- Split the url first and parse the query string
- Take the parameter out
- Rebuild the url
The below code is an example of such an algorithm:
// remove $qs_key from query string of $url
// return modified url value
function clean_url_qs($url, $qs_key)
{
// first split the url in two parts (at most)
$parts = explode('?', $url, 2);
// check whether query string is passed
if (isset($parts[1])) {
// parse the query string into $params
parse_str($parts[1], $params);
// unset if $params contains $qs_key
if (array_key_exists($qs_key, $params)) {
// remove key
unset($params[$qs_key]);
// rebuild the url
return $parts[0] .
(count($params) ? '?' . http_build_query($params) : '');
}
}
// no change required
return $url;
}
Test code:
echo clean_url('http://www.domain.com/myscript.php?p1=xyz&p2=10&p3=ghj', 'p2'), "\n";
echo clean_url('http://www.domain.com/myscript.php?p2=10&p3=ghj', 'p2'), "\n";
echo clean_url('http://www.domain.com/myscript.php?p2=10', 'p2'), "\n";