...but if I need to have a different @fields stored for each database?

Well, I told you the third solution with hashes is the best, because there undefined fields are not a problem at all. With strings you always have to set all fields. You just define the empty string (or a value that is not a valid value) for any field as an empty field. So if for example the dimension is undefined for a database you would have (AA,,N...). But using the empty string is only possible if it is not a valid value anywhere. If that is not the case you would have to use some ugly construct where '@@@' or something means undefined

It print only the fields of the "last" $Database acquired

'$Database' is not '$database'. upper/lower case is important.

I enclose the code, it's better, but I don't understand where you suggest to use an hash

 sub SubAcquisisciDatiCampiEsistenti {                    # Chiamata c
+on in INPUT $file,@Databases restituisce \%result
  my $file= shift;
  my @Databases= @_;
  my %result;
  foreach my $Database (@Databases) {
      my $cmd1 = "adarep db=$Database fdt file=$file ";
      #print "\n\n $cmd1 \n\n";
        my @ElencoDati= ();
        @ElencoDati = `$cmd1 |grep "  1       I  "`; 
        push (@ElencoDati, `$cmd1 |grep "SUPER"`);
        # my $ClearScr = `clear`;
        my @array1 = ();
        my @ElencoCampi = "";
        foreach (@ElencoDati) {
            @array1 = split /\s+/, $_;
                push (@ElencoCampi, "$array1[3],$array1[5],$array1[7],
+$array1[9]");
                }
        shift (@ElencoCampi);    
        @fields = sort (@ElencoCampi);
        chomp @fields;
        print "\n Elenco Campi Assegnati sul file $file del DB $Databa
+se Ordinati \n";
        foreach (@fields) {
        print "$_\n";
        }        
        print "\nL'ultimo campo utilizzato sul file $file del database
+ $Database e: ", $fields[$#fields],"\n\n\n";
        my @Caratteri = split (//,$fields[$#fields]);
        print "\n";
        my $PrimaLettera = ord $Caratteri[0];
        print "Il valore numerico nella tabella ASCII della prima lett
+era e: $PrimaLettera\n\n";
        my $SecondaLettera = ord $Caratteri[1];
        print "Il valore numerico nella tabella ASCII della seconda le
+ttera e: $SecondaLettera\n\n";
        print "\n\n";
        
        if ( $SecondaLettera < 90 ) {
            $SecondaLettera+=1;
            push (@nextfield, chr($PrimaLettera).chr($SecondaLettera))
+;
            # print "\n\nIl prossimo campo disponibile e: $ProssimoCam
+po\n\n";
            }    

        if ( $SecondaLettera == 90 ) {
            $PrimaLettera+=1;
            $SecondaLettera=65;
            push (@nextfield, chr($PrimaLettera).chr($SecondaLettera))
+;
            #$nextfield = chr($PrimaLettera).chr($SecondaLettera);
            # print "\n\nIl prossimo campo disponibile e: $ProssimoCam
+po\n\n";
            }
                
      print Dumper(\$Database,\@nextfield,\@fields);
       
       $result{$Database}{next}= \@nextfield;
    $result{$Database}{fields}= \@fields;
    }
  return \%result;
}
[download]

outside this sub looping to re-print the result of the printing give me only the last fields of the last array two times

In your original script you had a line "use strict;". Did you remove that line? It does seem that way, because otherwise the line "@fields = sort (@ElencoCampi);" would have given an error because you are using the global variable @fields instead of a local variable generated with my.

If you add a line "print \@fields;" to your loop you will see the problem, the reference never changes, it is always the same. But you need a new variable in every iteration of the loop so that they point to different memory locations

Change that line to "my @fields = sort (@ElencoCampi);" and your problem should go away.