Rather than shuffling the alphabet string , it is quicker to get a single random char .
Get a single random char from the string and then append the md5( time( ) )
to it . Before appending md5( time( ) )
remove one char from it so as to keep the resulting string length to 32 chars :
substr("abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ", mt_rand(0, 51), 1).substr(md5(time()), 1);
Lowercase version :
substr("abcdefghijklmnopqrstuvwxyz", mt_rand(0, 25), 1).substr(md5(time()), 1);
Or even shorter and a tiny bit faster lowercase version :
chr(mt_rand(97, 122)).substr(md5(time()), 1);
/* or */
chr(mt_rand(ord('a'), ord('z'))).substr(md5(time()), 1);
A note to anyone trying to generate many random strings within a second:
Since time( )
returns time in seconds , md5( time( ) )
will be same throughout a given second-of-time due to which if many random strings were generated within a second-of-time, those probably could end up having some duplicates .
I have tested using below code . This tests lower case version :
$num_of_tests = 100000;
$correct = $incorrect = 0;
for( $i = 0; $i < $num_of_tests; $i++ )
{
$rand_str = substr( "abcdefghijklmnopqrstuvwxyz" ,mt_rand( 0 ,25 ) ,1 ) .substr( md5( time( ) ) ,1 );
$first_char_of_rand_str = substr( $rand_str ,0 ,1 );
if( ord( $first_char_of_rand_str ) < ord( 'a' ) or ord( $first_char_of_rand_str ) > ord( 'z' ) )
{
$incorrect++;
echo $rand_str ,'<br>';
}
else
{
$correct++;
}
}
echo 'Correct: ' ,$correct ,' . Incorrect: ' ,$incorrect ,' . Total: ' ,( $correct + $incorrect );
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…