Unique email addresses by domain - php

Question

Im trying to make the below function only return 1 email per domain.
Example: if i feed the function:
email1#domain.com email2#domain.com email1#domain.com
email1#domain.com email3#test.co.uk
I want it to return
email1#domain.com email3#test.co.uk
Here is the current function:
function remove_duplicates($str) {
# match all email addresses using a regular expression and store them
# in an array called $results
preg_match_all("([\w-]+(?:\.[\w-]+)*#(?:[\w-]+\.)+[a-zA-Z]{2,7})",$str,$results);
# sort the results alphabetically
sort($results[0]);
# remove duplicate results by comparing it to the previous value
$prev="";
while(list($key,$val)=each($results[0])) {
if($val==$prev) unset($results[0][$key]);
else $prev=$val;
}
# process the array and return the remaining email addresses
$str = "";
foreach ($results[0] as $value) {
$str .= "<br />".$value;
}
return $str;
};
Any ideas how to achieve this?

Answer 1

Something along these lines:
$emails = array('email1#domain.com', 'email2#domain.com', 'email1#domain.com', 'email1#domain.com', 'email3#test.co.uk');
$grouped = array();
foreach ($emails as $email) {
preg_match('/(?<=#)[^#]+$/', $email, $match);
$grouped[$match[0]] = $email;
}
var_dump($grouped);
This keeps the last occurrence of a domain, it's not hard to modify to keep the first instead if you require it.

Answer 2

You could simply use the array_unique function to do the job for you:
$emails = explode(' ', $emailString);
$emails = array_unique($emails);

Answer 3

The concept prev is not reliable unless all equal hostnames are in one continuous sequence. It would work if you were sorting by hostname, with a sorting function provided, but it's a bit of overkill.
Build an array with the hostnames, drop entries for which there is already a hostname in the array.

Answer 4

I'd suggest the following trick/procedure:
Change from one string to array of addresses. You do this with preg_match_all, others might do it with explode, all seems valid. So you have this already.
Extract the domain from the address. You could do this again with an regular expression or some other thing, I'd say it's trivial.
Now check if the domain has been already used, and if not, pick that email address.
The last point can be easily done by using an array and the domain as key. You can then use isset to see if it is already in use.
Edit: As deceze opted for a similar answer (he overwrites the matches per domain), the following code-example is a little variation. As you have got string input, I considered to iterate over it step by step to spare the temporary array of addresses and to do the adress and domain parsing at once. To do that, you need to take care of the offsets, which is supported by preg_match. Something similar is actually possible with preg_match_all however, you would then have the array again.
This code will pick the first and ignore the other addresses per domain:
$str = 'email1#domain.com email2#domain.com email1#domain.com email1#domain.com email3#test.co.uk';
$addresses = array();
$pattern = '/[\w-]+(?:\.[\w-]+)*#((?:[\w-]+\.)+[a-zA-Z]{2,7})/';
$offset = 0;
while (preg_match($pattern, $str, $matches, PREG_OFFSET_CAPTURE, $offset)) {
list(list($address, $pos), list($domain)) = $matches;
isset($addresses[$domain]) || $addresses[$domain] = $address;
$offset = $pos + strlen($address);
}