• perldoc File::Find
• perldoc find2perl

This concludes today's RTFM on walking the filesystem. Tune in tomorrow, when Vivian confronts Alberto with the shocking news that his inode was actually fathered by Dirk.

use File::Basename qw/ dirname basename /;
use File::Find::Rule;

my $rule = rule( file => start => '/data/input' );

while(my $path = $rule->match) {
  printf "%s,%s,%s\n", dirname($path), basename($path), -s $path;
}
[download]

## do same 'use's as above
printf "%s,%s,%s\n", dirname($_), basename($_), -s $_
  for find( file => in => '/data/input' );
[download]

hello broquaint,

Thank you for your suggestion, unfortunately I don't have package File::Find. When I run your script, it gives me this error:

"go3.pl" [New file] 5 lines, 230 characters
normal@firestarter:/data/input> perl go3.pl
Can't locate File/Find/Rule.pm in @INC (@INC contains: /usr/perl5/5.00
+503/sun4-solaris /usr/perl5/5.00503 /usr/perl5/site_perl/5.005/sun4-s
+olaris /usr/perl5/site_perl/5.005 .) at go3.pl line 2.
BEGIN failed--compilation aborted at go3.pl line 2.
[download]

Then you need to install the module, and fortunately I've covered this very subject before at Re: Re: Re: Find file name. One thing I forgot to mention there was that you'll also need to install Number::Compare and Text::Glob, if you haven't already, to fulfil FFR's dependencies.
HTH

_________
broquaint

danield, it's worth persevering with the File::Find::Rule install. So much easier than the opendir route. If you have root, it's no harder than:

perl -MCPAN -e 'install File::Find::Rule';
[download]

-- vek --

This code keeps a array of directories and works through them sequentially. Finding a dir somewhere adds it to that array.
pelagic

I did check your suggestion, and the code now looks like this:


#!/usr/bin/perl -w
#!C:\Perl\bin\perl -w

my $directory='/data/input/data.01.29.2004-15:34:19/content';
my $size=0;
my $file;

#opendir(DIRHANDLE, $directory) or die "couldn't open $directory : $!\
+n";
#my@files = grep(!/^\.\.?$/,readdir(DIRHANDLE));


    push(@dirs,$directory);     # add an element to an array
    foreach $dir (@dirs)    # loop through the elements of an array
    {
        opendir(DIR,"$dir") || die "opendir $dir failed";
        my@list = grep(!/^\.\.?$/,readdir(DIR));
        closedir(DIR);
    
    foreach $file (@list)
      {
        $size=(-s "$dir/$file");
        print "$directory,$file,$size";
        print "\n";
      }
    };

#closedir(DIRHANDLE);
[download]


However it still does not do what I want. The script now gives a list of subdirectories of 'content' directory. Each subdirectory however contains files I am interested in. Example:

/data/input/data.01.29.2004-15:34:19/content/1/file001.txt
..
/data/input/data.01.29.2004-15:34:19/content/1/file100.txt
/data/input/data.01.29.2004-15:34:19/content/2/file001.txt
..
/data/input/data.01.29.2004-15:34:19/content/2/file100.txt
..

What I get is:

/data/input/data.01.29.2004-15:34:19/content,1,8192
/data/input/data.01.29.2004-15:34:19/content,2,8192

ok, you dropped an important part and you had a couple of typos also. Like that it works:

#!/usr/bin/perl -w
#!C:\Perl\bin\perl -w

my $directory='C:\dev\perl';
my $size=0;
my $file;

#opendir(DIRHANDLE, $directory) or die "couldn't open $directory : $!\
+n";
#my@files = grep(!/^\.\.?$/,readdir(DIRHANDLE));


push(@dirs,$directory);     # add  INITIAL  directory to an array
foreach $dir (@dirs)        # loop through the elements of an array
{
    opendir(DIR,"$dir") || die "opendir $dir failed";
    my@list = grep(!/^\.\.?$/,readdir(DIR));
    closedir(DIR);

    foreach $file (@list)
    {
        if (-d "$dir/$file") {
            push(@dirs, "$directory/$file");     # if this is a dir ad
+d it to an array
        }
        if (-f "$dir/$file") {            # only if it a file do your 
+size stuff
            $size=(-s "$dir/$file");
            print "$dir,$file,$size";
            print "\n";
        }
    }
};

#closedir(DIRHANDLE);
[download]


see comments in code ...
pelagic

But I'll add one more option (shameless plug): An alternative to File::Find -- I know, I know... lots of monks think it's dirty somehow to use an existing utility program in a subprocess to do something that a perl module can do. But "find" is available on all unix systems (including MacOSX, of course), it's easy to get an MS-windows port (find.exe) from numerous great sources, and (ahem) it's just faster and easier in all respects when compared to File::Find. Just my $0.02...