Beefy Boxes and Bandwidth Generously Provided by pair Networks
go ahead... be a heretic

Re: find biggest file and use awk

by graff (Chancellor)
on May 21, 2014 at 04:01 UTC ( #1086899=note: print w/replies, xml ) Need Help??

in reply to find biggest file and use awk

If all the files of interest are in one directory, and they all match a simple file-name pattern, and you just want to print the name of the largest file, here's one easy way to do that in Perl:
#!/usr/bin/perl use strict; use warnings; my $path = "/tmp/dir1"; my %filesize = map { $_ => -s } <$path/d1*>; my @sorted = sort { $filesize{$b} <=> $filesize{$a} } keys %files; printf "Largest file: %s (%d bytes)\n", $sorted[0], $filesize{$sorted[ +0]};
Of course, I expect it would be better to have the path and file-name pattern of interest be a command line argument (or a sensible default, like all files in the current working directory), because hard-coding this in the script is bothersome. So I'd rather do it like this:
#!/usr/bin/perl use strict; use warnings; my $glob_pattern = shift || './*'; my %files; for ( glob( $glob_pattern )) { $files{$_} = -s _ if ( -e ); } if ( scalar keys %files == 0 ) { warn "No files matched $glob_pattern\nUsage: $0 [path/name*]\n"; exit(1); } my @sorted = sort { $files{$b} <=> $files{$a} } keys %files; printf( "Largest file that matches %s is %s (%d bytes)\n", $glob_pattern, $sorted[0], $files{$sorted[0]} );
Now, in that case, when I run the script, I have to put the command-line argument in quotes, because otherwise, the shell will do the glob expansion, and my script will only see the first file name that matches the glob. In other words, if the script is called "show-biggest", the command line would have to be:
show-biggest '/tmp/dir1/d1*' # note the single quotes # or: show-biggest /tmp/dir1/d1\* # note the backslash escape for "*"
BTW, in trying this out, I learned that there is a subtle difference between this:
my @files = <something>;
and this:
my $glob = "something"; my @files = glob( $glob );
In the first approach, if "something" doesn't match anything, @files will be empty, but in the second approach, it will have one element, which is the string that was passed to the glob() function. The difference goes away if the value of $glob contains any wild-card characters (* or ? or square brackets) - I haven't checked, but I'll bet this is documented behavior... That's why I added a test for file existence (-e) in my second version of the script above.

I also learned that glob( $glob_pattern ) does the right thing, where <$glob_pattern> doesn't. (Perl treats the latter as an unopened file handle.)

Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://1086899]
and the fire pops...

How do I use this? | Other CB clients
Other Users?
Others musing on the Monastery: (6)
As of 2017-05-26 04:57 GMT
Find Nodes?
    Voting Booth?