How to make random string generator filter out duplicate strings?

Question

I have made a string generator:

    <?php

function createRandomPassword() {
    $chars = "abcdefghijkmnopqrstuvwxyz023456789";
    srand((double)microtime()*1000000);
    $i = 0;
    $pass = '' ;
    while ($i <= 12) {
        $num = rand() % 33;
        $tmp = substr($chars, $num, 1);
        $pass = $pass . $tmp;
        $i++;
    }
    return $pass;
}


// Usage
$product_names = array ();
for ($i=0; $i < 100; $i++ )
$product_names[] = "code-" . createRandomPassword();
?>

My problem is that I think there is a chance this could duplicate values, and its very important that doesnt happen as I will be generating about 700,000 of them.

Is there a good way to ensure the generated strings are not duplicates?

Thanks :)

Why go to all this trouble? Just use sha1() or md5(). As well, you do NOT re-seed the PRNG on each iteration of the loop. You seed it ONCE, and php does that for you automatically the first time you call rand() anyways. In practical terms, you'd only seed manually if you're testing and want the same 'random' stream each time. — Marc B
– Marc B, Commented Feb 27, 2012 at 14:47

chiborg · Accepted Answer · 2012-02-27 14:51:38Z

1

The following example will generate 100 product names, remove the duplicates and generate new product names until the product name array has a count of 100.

$cnt = 0;
$product_names = array ();
while($cnt < 100) {
   for ($i=0; $i < 100 - $cnt; $i++ )
     $product_names[] = "code-" . createRandomPassword();
   $product_names = array_unique($product_names);
   $cnt = count($product_names);
}

You can put the number 100 into a variable or constant and exchange it with 700,000.

edited Feb 27, 2012 at 14:51

answered Feb 27, 2012 at 14:45

chiborg

28.4k15 gold badges102 silver badges120 bronze badges

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

Comments

Joey · Accepted Answer · 2012-02-27 14:45:38Z

0

Several options:

Generate 700000 unique strings and shuffle them.
Use GUIDs

Although your code suggests they are used as passwords; if they really are, then don't bother about duplicates, as there are no obvious candidates that are possibly more likely (e.g. password or qwerty123).

answered Feb 27, 2012 at 14:45

Joey

357k88 gold badges704 silver badges699 bronze badges

Comments

Brad Christie · Accepted Answer · 2012-02-27 14:46:44Z

0

Create a look up/check table (in a database or file) or make an algorithm that makes a better unique key (using md5 or sha or something that's designed to be unique).

answered Feb 27, 2012 at 14:46

Brad Christie

102k16 gold badges160 silver badges200 bronze badges

2 Comments

chiborg Over a year ago

Some nitpicking here: SHA1 and MD5 are not designed to be collision-free, but have a very, very low probability of collision (which means they should be collision-free in practice).

Brad Christie Over a year ago

@chiborg: Understood, and well-received, however based on the OP's post, anything with [even] a 1% chance of collision is a vast improvement over using rand() and a fixed-string.

Collectives™ on Stack Overflow

How to make random string generator filter out duplicate strings?

3 Answers 3

Comments

Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related