natol44 has asked for the wisdom of the Perl Monks concerning the following question:

Hello,

Well after 2 days search, I give up and call for help :(

I have a list of files always on the same syntax:

string_amount_end

string is a 32 letters/digit chain (a md5 hash)
amount is always formatted with a digit and 2 decimals (1.00, 45.21, etc)
_end can be nul, or _min or _max
So a file name looks like 111aaa222ccc324567fed54333221235_1.02_max
or 111aaa222ccc324567fed54333221235_1.02
etc

I need to find all the files with the same "string" AND that have "_max" at the end, then inside those files I need to find the one where "amount" is the lowest.

So for example in the following list,:

111aaa222ccc324567fed54333221235_1.04
111aaa222ccc324567fed54333221235_1.05_max
111aaa222ccc324567fed54333221235_0.98_min
111aaa222ccc324567fed54333221235_1.02_max
111aaa222ccc324567fed54333221235_0.21
777aaa222ccc324567fed54333221235_1.04
777aaa222ccc324567fed54333221235_1.07_min
777aaa222ccc324567fed54333221235_1.04_max

if I am interested by the string 111aaa222ccc324567fed54333221235, the result should be 1.02 (from 111aaa222ccc324567fed54333221235_1.02_max)

Additionally I also need to know if there is no _max file at all in the selected files (that have the same "string").

Well I am a newbie, even a code of 20 lines didn't get the good result :(

Thank you for your help! (I add that it is not for business or exam etc, I am simply writing a game for my website :)

Replies are listed 'Best First'.
Re: File sort and values in file name
by ikegami (Patriarch) on Sep 18, 2009 at 19:39 UTC

    I need to find all the files with the same "string"

    Group by "string" ⇒ use a hash. 

    use List::Util qw( min );

my %grouped_maxes;
while (<DATA>) {
    chomp;
    my ($hash, $amount, $end) = split /_/;
    push @{ $grouped_maxes{$hash} }, $amount
        if defined($end) && $end eq 'max';
}

for my $hash (keys %grouped_maxes) {
    print("$hash: ", min( @{ $grouped_maxes{$hash} } ), "\n");
}

__DATA__
111aaa222ccc324567fed54333221235_1.04
111aaa222ccc324567fed54333221235_1.05_max
111aaa222ccc324567fed54333221235_0.98_min
111aaa222ccc324567fed54333221235_1.02_max
111aaa222ccc324567fed54333221235_0.21
777aaa222ccc324567fed54333221235_1.04
777aaa222ccc324567fed54333221235_1.07_min
777aaa222ccc324567fed54333221235_1.04_max
888aaa222ccc324567fed54333221235_0.21

    [download]

    Additionally I also need to know if there is no _max file at all in the selected files

    Keep track of all the unique hash strings you see, keep track of the unique hash strings you see with a max, then remove those with a max from the whole set.

    Remove duplicates ⇒ use a hash. 

    use List::Util qw( min );

my %hashes;
my %grouped_maxes;
while (<DATA>) {
    chomp;
    my ($hash, $amount, $end) = split /_/;
    ++$hashes{$hash};
    push @{ $grouped_maxes{$hash} }, $amount
        if defined($end) && $end eq 'max';
}

for my $hash (keys %hashes) {
    if (exists($grouped_maxes{$hash})) {
        my $min_max = min( @{ $grouped_maxes{$hash} } );
        print("$hash: $min_max\n");
    } else {
        print("$hash: No max\n");
    }
}

__DATA__
111aaa222ccc324567fed54333221235_1.04
111aaa222ccc324567fed54333221235_1.05_max
111aaa222ccc324567fed54333221235_0.98_min
111aaa222ccc324567fed54333221235_1.02_max
111aaa222ccc324567fed54333221235_0.21
777aaa222ccc324567fed54333221235_1.04
777aaa222ccc324567fed54333221235_1.07_min
777aaa222ccc324567fed54333221235_1.04_max
888aaa222ccc324567fed54333221235_0.21

    [download]
    777aaa222ccc324567fed54333221235: 1.04
888aaa222ccc324567fed54333221235: No max
111aaa222ccc324567fed54333221235: 1.02

    [download]
[reply]
[d/l]
[select]
Re: File sort and values in file name
by Bloodnok (Vicar) on Sep 18, 2009 at 20:08 UTC
    Please be advised that the norm is to include, if not the full, then a compilable/working sub-set of the, code that, when run, manifests the problem(s) you encountered on the way. Having said that... 
    use warnings;
use strict;

use Data::Dumper;

my @files = <DATA>;
chomp @files;

warn Dumper \@files;

sub get_least($) {
    my $md5 = shift;

    my $num = (
        sort map { s/${md5}_//; s/_max//; ($_) } grep /^$md5.*_max/, @
+files
    )[0];

    warn("No max file"), return undef unless $num;
    return $num;
}

warn get_least(q/111aaa222ccc324567fed54333221235/);
warn get_least(q/777aaa222ccc324567fed54333221235/);
warn get_least(q/99999999999999999999999999999999/);

__DATA__
111aaa222ccc324567fed54333221235_1.04
111aaa222ccc324567fed54333221235_1.05_max
111aaa222ccc324567fed54333221235_0.98_min
111aaa222ccc324567fed54333221235_1.02_max
111aaa222ccc324567fed54333221235_0.21
777aaa222ccc324567fed54333221235_1.04
777aaa222ccc324567fed54333221235_1.07_min
777aaa222ccc324567fed54333221235_1.04_max
99999999999999999999999999999999_1.04
99999999999999999999999999999999_1.07_min
99999999999999999999999999999999_1.04_min

    [download]
    when run, gives 
    $ perl tst.pl
$VAR1 = [
          '111aaa222ccc324567fed54333221235_1.04',
          '111aaa222ccc324567fed54333221235_1.05_max',
          '111aaa222ccc324567fed54333221235_0.98_min',
          '111aaa222ccc324567fed54333221235_1.02_max',
          '111aaa222ccc324567fed54333221235_0.21',
          '777aaa222ccc324567fed54333221235_1.04',
          '777aaa222ccc324567fed54333221235_1.07_min',
          '777aaa222ccc324567fed54333221235_1.04_max',
          '99999999999999999999999999999999_1.04',
          '99999999999999999999999999999999_1.07_min',
          '99999999999999999999999999999999_1.04_min'
        ];
1.02 at tst.pl line 22, <DATA> line 11.
1.04 at tst.pl line 23, <DATA> line 11.
No max file at tst.pl line 18, <DATA> line 11.
Use of uninitialized value in warn at tst.pl line 24, <DATA> line 11.
Warning: something's wrong at tst.pl line 24, <DATA> line 11.

    [download]
    which, I believe, is as required.

    A user level that continues to overstate my experience :-))
[reply]
[d/l]
[select]
Re: File sort and values in file name
by bichonfrise74 (Vicar) on Sep 18, 2009 at 23:01 UTC
    Try this... 
    #!/usr/bin/perl

use strict;

my %records;
while (<DATA>) {
    chomp;
    my @values = split( "_" );
    next unless $values[2] eq 'max';

    $records{ $values[0] } = $values[1]
      if ( $records{ $values[0] } > $values[1] ||
         ! $records{ $values[0] } );
}

print map { "$_ => $records{$_} \n" } keys %records;

__DATA__
111aaa222ccc324567fed54333221235_1.04
111aaa222ccc324567fed54333221235_1.05_max
111aaa222ccc324567fed54333221235_0.98_min
111aaa222ccc324567fed54333221235_1.02_max
111aaa222ccc324567fed54333221235_0.21
777aaa222ccc324567fed54333221235_1.04
777aaa222ccc324567fed54333221235_1.07_min
777aaa222ccc324567fed54333221235_1.04_max

    [download]
[reply]
[d/l]
      I liked your code, but I am averse to using [$index] in Perl code unless it is needed. I would give names to the list on the left as result of a split. A chomp() would make the split regex a bit easier, but no biggie. I am having a "brain cramp" right now on how to code the hash assignment statement better. My main point is to avoid this numeric $index stuff. 
      #!/usr/bin/perl -w
use strict;

my %records;
while (<DATA>) 
{
    my ($md5,$version,$end) = split(/[_\n]+/,$_);
    
    next unless ($end eq 'max');

    $records{$md5} = $version 
       unless (    exists $records{$md5} 
                && $records{$md5}< $version) ;                
}

foreach my $md5 (sort keys %records)
{
   print $md5,"_",$records{$md5},"_max\n";
}


#Prints:
#111aaa222ccc324567fed54333221235_1.02_max
#777aaa222ccc324567fed54333221235_1.04_max


__DATA__
111aaa222ccc324567fed54333221235_1.04
111aaa222ccc324567fed54333221235_1.05_max
111aaa222ccc324567fed54333221235_0.98_min
111aaa222ccc324567fed54333221235_1.02_max
111aaa222ccc324567fed54333221235_0.21
777aaa222ccc324567fed54333221235_1.04
777aaa222ccc324567fed54333221235_1.07_min
777aaa222ccc324567fed54333221235_1.04_max

      [download]
[reply]
[d/l]
[select]

Back to Seekers of Perl Wisdom