Swear filter case sensitive - php

Question

I have a little problem with my function:
function swear_filter($string){
$search = array(
'bad-word',
);
$replace = array(
'****',
);
return preg_replace($search , $replace, $string);
}
It should transform "bad-word" to "**" but the problem is the case sensivity
eg. if the user type "baD-word" it doesn't work.

Answer 1

The values in your $search array are not regular expressions.
First, fix that:
$search = array(
'/bad-word/',
);
Then, you can apply the i flag for case-insensitivity:
$search = array(
'/bad-word/i',
);
You don't need the g flag to match globally (i.e. more than once each) because preg_replace will handle that for you.
However, you could probably do with using the word boundary metacharacter \b to avoid matching your "bad-word" string inside another word. This may have consequences on how you form your list of "bad words".
$search = array(
'/\bbad-word\b/i',
);
Live demo.
If you don't want to pollute $search with these implementation details, then you can do the same thing a bit more programmatically:
$search = array_map(
create_function('$str', 'return "/\b" . preg_quote($str, "/") . "\b/i";'),
$search
);
(I've not used the recent PHP lambda syntax because codepad doesn't support it; look it up if you are interested!)
Live demo.
Update Full code:
function swear_filter($string){
$search = array(
'bad-word',
);
$replace = array(
'****',
);
// regex-ise input
$search = array_map(
create_function('$str', 'return "/\b" . preg_quote($str, "/") . "\b/i";'),
$search
);
return preg_replace($search, $replace, $string);
}

Answer 2

I think you mean
'/bad-word/i',

Answer 3

Do you even need to use regex?
function swear_filter($string){
$search = array(
'bad-word',
);
if (in_array(strtolower($string), $search){
return '****';
}
return $search
}
makes the following assumptions.
1) $string contains characters acceptable in the current local
2) all contents of the $search array are lowercase
edit: 3) Entire string consists of bad word
I suppose this would only work if the string was split and evaluated on a per word basis.