How to produce a unique string from a php array

Question

I need a unique string from an array so that I can tell when it changes without measuring the inputs of that array. I'm trying to work out if it is computationally efficient to calculate a value rather than add code to look out for changes in the array. The array itself can have a variety of values and for future proofing I don't want to try and measure whether new values have been added to the array, I'd much rather just create some string or hash that will change if the array itself changes.

So for example:

$a = Array(
'var1' => 1,
'var2' => 2,
'var3' => 3,
);

If I was to use md5(http_build_query($a)) perhaps with an added ksort to confirm that the order of the keys haven't changed this might then produce a unique string that I can use to compare against another run of the application to evaluate whether the array has changed.

I'm looking for an alternate, possibly faster or more elegant solutions to this.

I don't think array_diff would deal with checking if the key order has changed. You could json_encode the array and take a hash of that. Might want to check the performance of json_encode and http_build_query. — Endophage
– Endophage, Commented Feb 25, 2011 at 0:33

delphist · Accepted Answer · 2011-02-25 00:28:10Z

8

Im use md5(serialize($array)) for this. Its better, because works for multi-dimensional arrays.

answered Feb 25, 2011 at 0:28

delphist

4,5591 gold badge24 silver badges22 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Jason · Accepted Answer · 2011-02-25 00:35:24Z

3

Thanks for all the ideas guys.

I've tried all of them except a sha-256 which my server doesn't have installed.

Here's the results:

Average (http_build_query): 1.3954045954045E-5
Average (diff): 0.00011533766233766
Average (serialize): 1.7588411588412E-5
Average (md5): 1.6036963036966E-5
Average (implode-haval160,4): 1.5349650349649E-5

That's running the operation 1000 times and averaging the result. After refreshing a couple times I could tell that the http_build_query was the quickest. I guess my next question would be if anyone can think of any pitfalls of using this method?

Thanks

Here's my code:

class a {

    static $input;

    function test() {
        $start = null;
        $s = $e = $d = $g = $h = $i = $k = array();
        self::$input = array();

        for ($x = 0; $x <= 30; $x++) {
            self::$input['variable_' . $x] = rand();
        }

        for ($x = 0; $x <= 1000; $x++) {
            $start = microtime();

            $c = http_build_query(self::$input);
            ($c == $c);

            $s[] = microtime() - $start;
        }

        for ($x = 0; $x <= 1000; $x++) {
            $start = microtime();

            $c = md5(http_build_query(self::$input));
            ($c == $c);

            $e[] = microtime() - $start;
        }

        for ($x = 0; $x <= 1000; $x++) {
            $start = microtime();

            $c = array_diff(self::$input, self::$input);

            $d[] = microtime() - $start;
        }
        for ($x = 0; $x <= 1000; $x++) {
            $start = microtime();

            $c = serialize(self::$input);
            ($c == $c);

            $g[] = microtime() - $start;
        }

        for ($x = 0; $x <= 1000; $x++) {
            $start = microtime();

            $c =  hash("haval160,4", implode(',',self::$input));
            ($c == $c);

            $h[] = microtime() - $start;
        }
        echo "<pre>";

//print_r($s);
        echo "Average (http_build_query): " . array_sum($s) / count($s) . "<br>";
        echo "Average (diff): " . array_sum($d) / count($d) . "<br>";
        echo "Average (serialize): " . array_sum($g) / count($g) . "<br>";
        echo "Average (md5): " . array_sum($e) / count($e). "<br>";
        echo "Average (implode-haval160,4): " . array_sum($h) / count($h);
    }

}

a::test();

answered Feb 25, 2011 at 0:35

Jason

15.4k15 gold badges71 silver badges106 bronze badges

2 Comments

delphist Over a year ago

damn its genious array_diff(self::$input, self::$input);

Jason Over a year ago

well in practical use, for this application, most of the time it will be a match so I decided that I wanted to count the time it takes to identify a match rather than identify a non match. If I am correct, it would take less time to identify a non match because it has to iterate and check every key. It would only stop checking if it found a difference, which it doesnt, so going by that logic hopefully that means I am being more conservative by comparing to of the exact same variables

krtek · Accepted Answer · 2011-02-25 00:31:12Z

2

PHP has an array_diff() function, don't know if it's of any use for you.

Otherwise, you can eventualy use the incremental hashing possibility offered by php : http://www.php.net/manual/en/function.hash-init.php by iterating over each values of the array and adding them in the incremental hash.

answered Feb 25, 2011 at 0:31

krtek

26.7k5 gold badges60 silver badges84 bronze badges

Comments

shmeeps · Accepted Answer · 2011-02-25 00:27:14Z

1

You could always just do

$str = implode(",", $a);
$check = hash("sha-256", $str);

Theoretically, that should detect changes in array size, data, or ordering.

Of course, you can use whatever hash you wish.

answered Feb 25, 2011 at 0:27

shmeeps

7,8033 gold badges31 silver badges35 bronze badges

1 Comment

shmeeps Over a year ago

Actually, delphist is probably correct, serialize() would probably work better if you didn't know the dimension of the array.

Collectives™ on Stack Overflow

How to produce a unique string from a php array

4 Answers 4

Comments

2 Comments

Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

Comments

2 Comments

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related