++bobf; I often used that line, but recently changed to using this form of it:

next if not /\S/; # Skip blank lines
[download]

With my brain under the influence of Perl Best Practices, I see some possible bugs, and opportunities for improvements.

• %saldi is not getting cleared between files! Is this what you want?
• Magic <> is nice, but when you need eof() processing over multiple files, consider changing to do your own opens/closes, etc.
• Sort your hash keys, or your output order may vary between runs.
• Your totals will not line up when printed; consider printf() to format the output.
• The code will be clearer if you name your fields instead of treating them as numbers, especially since the numbers change! (fields 3,4 become 0,1)
• You are not chomping your input, so your last field contains newlines. It has not caused problems because the last field is only handled numerically.

use strict;
use warnings;

foreach my $filename (@ARGV) {
    open my $fh, '<', $filename
      or  warn "Can't open $filename: $!"
      and next;

    my %saldi;
    while ( <$fh> ) {
        chomp;
        next if not /\S/; # Skip blank lines

        my @fields = split /,/;
        if ( @fields != 5 ) {
            warn "Bad number of fields in file '$filename' line $.";
            next;
        }

        my ( $department, $currency, $year, $type, $amount ) = @fields
+;
        $saldi{ $type } += $amount;
    }

    my ($basename) = ( $filename =~ m/^(\S+)\.txt/ )
      or warn;
    
    print "$basename\n";
    foreach my $type ( sort keys %saldi ) {
        my $total_amount = $saldi{$type};
        printf "\t%-7s\t%7.2f\n", $type, $total_amount;
    }

    close $fh
      or warn "Can't close $filename: $!";
}
[download]

Sure - sorry for being too cryptic. Sometimes it's hard to tell how much detail is needed. :-)

Insert a new line as follows (code snippet taken from your OP):

while (<>) {
    next if m/^\s+$/; # <-- add this line
    my @cellen = ( split /,/, )[ 3, 4 ];
[download]

The regex is anchored to the start (^) and end ($) of the string, and the pattern consists of one or more white space characters (\s+). See perlre and Pattern Matching, Regular Expressions, and Parsing (in the Tutorials section) for more information.

Incidently, the code I added is a bit shorter than it could be, since it utilizes $_. A more verbose version could be:

next if $_ =~ m/^\s+$/;
[download]

or, if you wanted to explicitly assign the input text to a variable before processing, you could do something like this:

while ( my $line = <>) {
    next if $line =~ m/^\s+$/; # <-- add this line
    my @cellen = ( split( /,/, $line ) )[ 3, 4 ];
[download]

HTH