Re: Re^3: Efficient processing of large directory

[BrowserUk]

in reply to Re^3: Efficient processing of large directory
in thread Efficient processing of large directory

readdir is less efficient than glob if a subset of the directory contents is sought.

[Aristotle]

Apart from the fact that the OP isn't looking for a subset.

#!/usr/bin/perl
use strict;
use warnings;
use Benchmark qw /cmpthese/;

chdir shift or die;

cmpthese -5 => {
        glob => sub {
                1 while glob '*';
        },
        readdir => sub {
                opendir my $dh, '.' or die;
                1 while readdir $dh;
        },
};
__END__

$ mkdir x ; cd x ; touch `seq 1 20000` ; cd ..
$ perl glob_vs_readdir.pl x
Benchmark: running glob, readdir for at least 5 CPU seconds...
   glob:  6 wallclock secs ( 4.02 usr +  1.07 sys =  5.09 CPU) @  5.30
+/s (n=27)
readdir:  6 wallclock secs ( 3.93 usr +  1.37 sys =  5.30 CPU) @ 65.28
+/s (n=346)
          Rate    glob readdir
glob    5.30/s      --    -92%
readdir 65.3/s   1131%      --