At 11:41 AM 8/6/01 -0400, Chris Rogers wrote:
>If your data is already in the format shown below as yearmonthday, then a
>standard sort would do just fine:
>
>@test =(20010327, 20001113, 20011225, 19991231, 20100115);
>foreach (sort @test)
Doh, I should have realized that this data also sorts asciibetically as
well as numerically, so I could have left the '{ $a <=> $b }' out of my
posting.
Compelled to redeem myself, I decided to see whether the numeric sort might
have an advantage, i.e., performance. Fortunately, it does:
use Benchmark 'cmpthese';
my (@unsorted, @sorted);
for (1 .. 200000) {
push @unsorted, sprintf "%04d%02d%02d", rand(10000), rand(100), rand(100);
}
cmpthese (10 =>
{ ascii => sub { @sorted = sort @unsorted },
numeric => sub { @sorted = sort { $a <=> $b } @unsorted }
});
Benchmark: timing 10 iterations of alpha, numeric...
ascii: 85 wallclock secs (84.36 usr + 0.27 sys = 84.63 CPU)
@ 0.12/s (n=10)
numeric: 80 wallclock secs (79.70 usr + 0.25 sys = 79.95 CPU)
@ 0.13/s (n=10)
s/iter alpha numeric
ascii 8.46 -- -6%
numeric 8.00 6% --
Thin justification, I know :-)
--
Peter Scott
Pacific Systems Design Technologies
http://www.perldebugged.com
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
