You should be able to do the whole thing including file_C in one pass of file_A.

#!perl -slw
use strict;
use vars qw[$NEW_A];
use Data::Dumper;
use Inline::Files '-backup';
local $,=', ';

my %b = map{ (
                '1'.$_->[0] => 1,
                '4'.$_->[1] => 1,
        ) }
        map{ [ (split':')[0,3] ] } <FILE_B>;

my %c = map{ $_ => 1 } map{ (split':')[2] } <FILE_C>;

#print Dumper \%b, \%c;

open NEW_A, "> $NEW_A" or die $!;

while(<FILE_A>) {
    chomp;
    my @fields = split':';
    print( "Removing '$_' from FILE_A because field 1:'$fields[0]' not
+ in FILE_B"),
        next if !exists $b{ '1'.$fields[0] };
    print( "Removing '$_' from FILE_A because field 4:'$fields[3]' not
+ in FILE_B"),
        next if !exists $b{ '4'.$fields[3] };
    print( "Removing '$_' from FILE_A because field 3:'$fields[2]' not
+ in FILE_C"),
        next if !exists $c{ $fields[2] };
    print NEW_A $_, $/;
}

__FILE_A__
1:*:Z:A:*
2:*:Y:B:*
3:*:X:C:*
4:*:Z:D:*
5:*:Y:E:*
6:*:X:F:*
7:*:T:G:*
__FILE_B__
1:*:*:A:*
2:*:*:B:*
3:*:*:C:*
4:*:*:D:*
5:*:*:E:*
6:*:*:F:*
__FILE_C__
*:*:Z:*:*
*:*:Y:*:*
*:*:T:*:*
__NEW_A__
1:*:Z:A:*
2:*:Y:B:*
4:*:Z:D:*
5:*:Y:E:*
[download]

open A, "<fileA" || die $!;
open B, "<fileA" || die $!;
open OUT, ">fileAtmp" || die $!;
my ($bufA, $bufB);
while ($bufA = <A>) {
  $bufB = <B> || last; #jump out if file B is shorter than A
  my @A = split ':', $bufA;
  my @B = split ':', $bufB;
  print OUT $bufA if $A[0] eq $B[0];
}
close A;
close B;
close OUT;
[download]

open A, "<fileA" || die $!;

Aherm, ITYM or not ||. Precedence matters.

I'm kinda confused as to what you did, so I'm going to try to write a translation of what you say in Perl:

my $file_A = read_file( "file_A" );
my $file_B = read_file( "file_B" );

foreach (keys %$file_A) {
  delete $file_A->{ $_ } unless exists( $file_B->{ $_ } );
}

sub read_file {
  my $file = shift;
  open FILE, $file or die "Can't open '$file' for reading $!";
  my @file = <FILE>;
  close FILE;

  my $hoa = {};
  foreach ( @file ) {
    my (@fields) = split /:/;
    my $key = $fields[0];
    $hoa->{ $key } = [ @fields ];
  }
  return $hoa;
}
[download]

If this is how you parsed the files, then it seems to me you could easily do the following:

sub change_field_to_check {
  my $hoa = shift;
  my $field = shift;
  $field -= 1;

  my $new_arrangement;

  foreach (keys %$hoa) {
    $new_arrangement->{ $hoa->{ $_ }[ $field ] } = $hoa->{ $_ };
  }

  return $new_arrangement;
}

$file_A = change_field_to_check( $file_A, 4 ); # Field 4
$file_B = change_field_to_check( $file_B, 4 ); # Field 4

foreach (keys %$file_A) {
  delete $file_A->{ $_ } unless exists( $file_B->{ $_ } );
}
[download]

This ought to extrapolate to other files and other fields. If your fields are named something else, having a hash that translates the field name to their order will give you the right key.

Hope this helps.

- m.

If order of lines in file_B is not important, and if there's no one-to-one line relationship order-wise between file_A and file_B, I'd change your algorithm to this:

• Read file_B and hash each line twice (two hashes): a hash for field_1 and a hash for field_4. Hash values are just 1 (true) for "found it in file_B".
• Read file_C and hash its field_3 the same way.
• Now open file_A and an output file of the desired lines. For each line in file_A, split the line. Now you just compare the split values to each hash and decide whether to write it to the temporary file or not. Write the lines you want to keep to the temporary file.

use Tie::File;
use strict;
my (@A, @B, @C); my ($field1, $A, $B);

tie @B, 'Tie::File', 'txtA.txt' or die "couldn't TiE $!\n";
tie @A, 'Tie::File', 'txtB.txt' or die "couldn't TiE $!\n";

foreach $A (@A) {
    my ($field1) = split/:/, $A;
    print $field1;

foreach $B (@B){
    if($B=~/^$field1/){ 
    #found match push it onto C
    push @C, $A;}
  }
}
## We remove those that didn't match by replacing @A with @C :)
@A = @C;
print @A;
[download]

filea:file2:file1
file2:something:old
[download]

fileB:file2:file3
filea:file2:file1
file2:filex:file23
file34:file2:file3
filea:file2:file1
file2:filex:file23
[download]