Unique Random Values II

This second method to get 3 random values off an array (with duplicate values) in PHP was actually borne out of a hunch that there's probably a more efficient way to get this done! And I was right...

Searching an array randomly and return 3 unique values - Method II

Using the not-so-popular PHP in_array() function (at least, not to me), here's how I re-wrote the script.

<?php
// replace this with the previous piece of code

for( $i=0; $i<3; $i++ ) // just get 3 diff. sites
{
  $r = mt_rand( 0, sizeof($sites)-1 ); // generate random key

  //  METHOD II
  //  =========
  if( isset($trio) ) # var $trio will hold the lucky 3 sites
  {
    if( in_array($sites["$r"], $trio) )
    {
      --$i;
    }
    else
    {
     $trio[] = $sites["$r"];
    }
  }
  else
  {
    $trio[] = $sites["$r"];
  }
  //  END of METHOD II
  //  ================
}

?>

Method I vs. Method II

While both methods fared well when I was looking to get just 3 random and unique values, the larger this number, the slower Method I gets...

I tried giving either script 1,000 different values in an array and for it to return 999 unique values and here's the result:

Method I        |Method II
----------------+-----------------
> 22 seconds!   |1.2 - 1.4 seconds
                |

It's quite impressive how the speed of Method II (i.e. using the in_array() function) just increases fractionally with so much more data to process!