while (<F1>)
   {
      /(\w*) (\d*)/ ;
      $hash1{$1} = $2 ;
   }
[download]

This won't solve your problem, but it's worth pointing out that you should never use $1, $2, etc without checking that the match succeeded. If the match failed then you'll get the values from the previous successful match.

while (<F1>) {
  if (/(\w*) (\d*)/) {
    $hash1{$1} = $2;
  } else {
    warn "Invalid input line $.: $_\n";
  }
}
[download]

Thanks.
This is not something I'd thought of but explains some weird results I've seen

Hi, If i understood your question correctly, this will help you.

open F1,">>F1.dat" or die "Can't open F1.dat: $!\n"; #Open F1 file in 
+Append mode
open F2,"F2.dat" or die "Can't open F2.dat: $!\n";
while(<F2>)
{
    print F1 $_;
}
close(F1);
close(F2);
[download]

(Untested)

My code creates the new file so I don't get duplicates and also deals with orphaned records i.e. those that only live in one file but not the other
Hence this doesn't help. Thanks anyway

You might want to look again at Samy_rio's solution above -- with a little tweaking, you could address the duplicate issue. I don't think that you will have a problem with orphaned records using the proposed file concatenation.

If you open the first file, step through it and create a lookup hash, then close and re-open the first file for appending and the second file for reading, you can simply check each line of the second file to see if it is already in your lookup hash before appending it to the first file.

You don't tell us what impact the additional words have in terms of the logic of your program -- if they are simply extra fields that you want to bring forward, then you can simply adjust your existing regular expression from (\d*) to (\d+.*) to include everything to the end of the line, something like this:

  if (/^(\w*) (\d+.*)$/) {
    unless (exists($lookup{$1}) && ($lookup{$1} eq $2)) {
      print F1, $_;
    }
  }
  } else {
    warn "Invalid input line $.: $_\n";
  }
[download]

If the additional words in the line have some impact on your program's logic (e.g., if you need to parse them off and possibly create new records in your resulting file) then you need to tell us.

Hope that helps. :)

Update: added 'exists' logic to check in lookup hash.


No good deed goes unpunished. -- (attributed to) Oscar Wilde

#!/usr/bin/perl -w
use strict;
my %d = map{
        /^(\w+) /; 
        $1, $_;
    }grep{
        chomp;
        /^\w+ \d+/;
    }map{
        local @ARGV = ($_);
        <>;
    }<F*.dat>;
open(MRG, ">merge.dat") or die "Oops! There was a problem: $!";
print MRG, $d{$_}."\n" for(sort keys %d);
close MRG;
[download]

if ($\$loop ne $NULL) { $hash1{$\$loop} = $\$loop+1 ; }
[download]

Since you are merging the hashes, do you really need two of them? Also, it seems that the only critical part of the line is the word at the begining, so you can just split that off separately and save the rest of the line in a single variable.

use warnings;
use strict;

my %filehash;

my @files = qw(F1.dat F2.dat);

for my $filename (@files) {
    open my $filehandle, '<', $filename or die "Can't open $filename $
+!\n";
    while (<$filehandle>) {
        my ( $word, $remainder ) = split ' ', $_, 2;
        $filehash{$word} = $remainder;
    }
    close $filehandle;
}

open my $merged, '>', 'merge.dat' or die "Can't open merge.dat:$!\n";
for ( sort keys %filehash ) {
    print $merged "$_ $filehash{$_}";
}
[download]

Update: changed split parameter to ignore any initial whitespace