Question

My website use various resources from a single domain, for example:

http://static.example.com/javascript/common.js
http://static.example.com/javascript/common.css
http://static.example.com/javascript/menu/menu.js
http://static.example.com/javascript/menu/menu.css
http://static.example.com/images/1804/logo/02000100.jpg
http://static.example.com/images/1804/headers/main/09400060.png
http://static.example.com/images/1804/headers/home/1101/06900200-01.jpg
http://static.example.com/images/1804/headers/home/1101/06900200-02.jpg

I need a very simple string hashing function that maps these URLs to numbers, the numbers being 0, 1, 2 and 3. The algorithm should be deterministic and uniform. I have tagged the question PHP but a generic answer is acceptable.

You might have guessed why I need this; I plan to change the URLs to, for example:

http://0.static.example.com/javascript/common.js
http://2.static.example.com/javascript/common.css

Answer 1

I prefer doing a crc32 hash of the string, and taking its modulo with the limit.

Code:

function numeric_hash($str, $range) {
    return sprintf("%u", crc32($str)) % $range;
}

Usage:

$str = "http://static.example.com/javascript/common.js
http://static.example.com/javascript/common.css
http://static.example.com/javascript/menu/menu.js
http://static.example.com/javascript/menu/menu.css
http://static.example.com/images/1804/logo/02000100.jpg
http://static.example.com/images/1804/headers/main/09400060.png
http://static.example.com/images/1804/headers/home/1101/06900200-01.jpg
http://static.example.com/images/1804/headers/home/1101/06900200-02.jpg";
$urls = explode("\n", $str);

foreach($urls as $url) {
    echo numeric_hash($url, 4) . "\n";
}

Output:

1
3
3
3
1
3
1
3