Re: Randomise a string of letters

[BrowserUk]

in reply to Randomise a string of letters

This F-Y shuffles the string in-place avoiding the split and join. It might be a bit more efficient, but I haven't benchmarked it.

#! perl -slw
use strict;

sub shuffleStr {
    my $len = length $_[0];
    my ($tmp, $n);
    $n = $_+rand($len-$_)
    , $tmp = substr( $_[0], $_, 1)
    , substr( $_[0], $_, 1) = substr( $_[0], $n , 1)
    , substr( $_[0], $n , 1) = $tmp
        for 0 .. $len;
    $_[0];
}
my $string = 'the quick brown fox';

print shuffleStr($string) for 1..10;
__END__
C:\test>227145
wconbex ftkuoiqhr
xhe kbquio nfwo tcr
xen iu kwoq rtobhcf
h ftcx wikboqnreou
eoqkxhitfwun  ocrb
tbikr nuoqofwxh c e
rhtwi efuc kxobnq o
noit oquxhke bwcfr
qruwoox  eftnih kcb
oo qwnfet bcxurih k

C:\test>

---

Examine what is said, not who speaks.

The 7th Rule of perl club is -- pearl clubs are easily damaged. Use a diamond club instead.

[matth]

Does anyone know how all these approaches compare with the EMBOSS program shuffleseq?

[BrowserUk]

I don't know for sure, but I assume (on the basis of what you have said in other threads) that EMBOSS is a gene sequence manipulation program/suite.

Given that genes sequences tend to be represented by pairs of chars, it quite probable that the routine you mention shuffles pairs of bytes around.

It's also quite likely, though again I don't really know, that this would be done in C, which is probably much faster that any implementation in pure perl. Though it's possible that one of the modules above implements the shuffle using XS or C in which case they could be comaparable for performance.

In terms of "true randomness", provided the shuffle is reasonable well written, the randomness will come down to the randomness of the RNGenerator, and unless your involved in crytoanalysis or heavy statistical analyisis, the RNG on most systems is good enough for most purposes.

---

Examine what is said, not who speaks.

The 7th Rule of perl club is -- pearl clubs are easily damaged. Use a diamond club instead.