By: Jeremy Cook

Jeremy Cook — Mon, 12 Jul 2010 14:25:29 +0000

Thanks Keith, that one is faster than ‘Hamlet’ but it didn’t work for me under Apache on Windows. If my theory is correct the problem is the use of str_shuffle(), which generated huge numbers of duplicates for me after just 4 iterations.

By: Keith Evetts

Keith Evetts — Mon, 12 Jul 2010 09:19:41 +0000

I should add that on Apache, the following method is fifty times quicker than encrypting Hamlet (you need to write an Exception class to go with it):

function get_codes( $numcodes, $codelength , $wastage = 25 ) {
if ( is_int($numcodes)
&& is_int($codelength)
&& $codelength >= 5
&& $codelength <= 40
&& is_int($wastage) )
{

// acceptable characters
$goodchars = '2346789AaBbDdEeFfHhJjKkMmNnPpRrTtWwXxYyZz';
// build array allowing for possible wastage through duplicate values
do { $unique_array[] = substr ( str_shuffle( $goodchars ) , 0 , $codelength );
} while ( sizeof ( $unique_array ) $numcodes + 1 ) {
return array_slice ( $unique_array, 0, ($numcodes + 1) );
} else throw new CodeGenException ( $numcodes – $size.’ fewer codes produced than requested: wastage’, 1, $numcodes, $codelength );
} else throw new CodeGenException ( ‘function get_codes called with wrong params’, 2, $numcodes, $codelength );
} // end func
___________________________
Notes:
- the string $goodchars comprises alphanumeric characters optimised to reduce visual confusions
- $goodchars must be longer than max codelength. Currently 41 chars.
- $wastage is tuned for default of 25 for a 5-character code
- average time to execute for an array of 50000 codes of 8-character length is 0.25 seconds
- you can try…catch CodeGenException # 1 and try the function again with a greater allowance for wastage

Comments on: Creating 50,000 unique values

By: Jeremy Cook

By: Keith Evetts