Hello rahu_6697,

Welcome to the Monastery. From my point of view I would approach the problem a bit differently. I would use stat to get each file latest updated date, sample:

my $modtime = (stat($fh))[9]
[download]

As a second step I would create a hash of hashes. The name of the file I would use it as key then the secondary keys would be keywords like stat and path or location. Sample of hash of hashes (not tested):

my %HoH = ( file_name => {
    stat  => 'date here',
    location => 'path to file'
    );
[download]

By iterating through the directories with your files you can check one by one the file and in case of common name you can compare the latest date and either update the hash or move ot the next one.

There are many modules that can help you find the file by searching the directories recursively for example File::Find

I would approach the problem like this.

Hope this helps. In case that something is not clear let me know and I will try to analyse my answer further.

BR, Thanos

The glob function has issues, see To glob or not to glob. But if your pattern is always going to be the fixed string you showed, without variables interpolated into the string, then it's probably ok in this case.

When you say "files with unique names", I'm going to assume you mean basenames (the filename without the path), but I'm also guessing that what you need in @filelist is still the full filename, so that you can properly open the files. Also, by "latest updated/created", I'm guessing you mean the mtime (see also stat(2)).

You could use a hash; here I'm using the basename as the key, and each value is an anonymous array where the first element is the full filename and the second element is the mtime. To see how it works, uncomment the "# Debug" lines.

use warnings;
use strict;
#use Data::Dump; # Debug
use File::Basename qw/fileparse/;
use File::stat;

my %filelist;
for my $file (glob '/user/*/ws/*/BLK_*/*.txt') {
    my $bn = fileparse($file);
    my $mtime = stat($file)->mtime;
    #dd $file, $bn, $mtime; # Debug
    $filelist{$bn} = [$file,$mtime]
        unless $filelist{$bn} && $filelist{$bn}[1]>$mtime;
}
#dd \%filelist; # Debug
my @filelist = sort map {$_->[0]} values %filelist;
#dd \@filelist; # Debug
[download]

Thanks for help, you have understood the problem quite well, I have to sort files with same basename present in different directory and subdirectories according to their mtime such that only latest modified file would come in @filelist. Later I have to work on these files separately to fetch some particular data which is working fine for me.I have done changes as per your code but on terminal these errors are coming. Please resolve this issue as I am working on a shared workspace so I can't install any package in linux.

Use of uninitialized value $arg in stat at /user/s5e7420/local/lib/per
+l5/5.18.0/File/stat.pm line 206.
Use of uninitialized value $name in index at /user/s5e7420/local/lib/p
+erl5/5.18.0/Symbol.pm line 118.
Use of uninitialized value $name in pattern match (m//) at /user/s5e74
+20/local/lib/perl5/5.18.0/Symbol.pm line 121.
Use of uninitialized value $name in hash element at /user/s5e7420/loca
+l/lib/perl5/5.18.0/Symbol.pm line 121.
Use of uninitialized value $name in concatenation (.) or string at /us
+er/s5e7420/local/lib/perl5/5.18.0/Symbol.pm line 129.
Can't call method "mtime" on an undefined value at script.pl line 13.
[download]

Hello again rahu_6697,

Fellow Monk haukex have provided a full answer to your question, but I just wanted to add something here. You wrote: Please resolve this issue as I am working on a shared workspace so I can't install any package in linux.

Programming and troubleshooting is part of the learning curve. Even if the fellow Monk haukex will solve the problem that you have today for you what you have learned? He also gave you a hint in the sample of code Data::Dumper module is for debugging purposes. Did you tried to debug your code? It is part of fun/hell debugging and troubleshooting.

Regarding the part that you are not able to install modules, Data::Dumper, File::stat and File::Find are core modules.

#!/usr/bin/env perl
use strict;
use warnings;
use Module::CoreList;

my @moduleList = ("Data::Dumper", "File::stat", "File::Find");

foreach my $module (@moduleList) {
    if (Module::CoreList::is_core($module)) {
    print "".$module." is a core module and it was released:";
    print Module::CoreList->first_release($module) . "\n";
    }
    else {
    print "".$module." it is not core module!\n";
    }
}

__END__

$ perl test.pl
Data::Dumper is a core module and it was released:5.005
File::stat is a core module and it was released:5.004
File::Find is a core module and it was released:5
[download]

Hope this helps, BR.

Seeking for Perl wisdom...on the process of learning...not there...yet!

Use of uninitialized value $arg in stat at /user/s5e7420/local/lib/perl5/5.18.0/File/stat.pm line 206.

It looks like you're passing undef to stat, but I'm not aware of a situation where glob (Update: in list context) might return undef. Could you show the minimal code needed to reproduce this error? (Short, Self-Contained, Correct Example, How do I post a question effectively?, and I know what I mean. Why don't you?)

Mtime prints the name and modification time (in seconds since the epoch) of each of the files.

HTH

--Chris