Try adding use diagnostics; at the top of your program to get explanations of the messages. In this case it's just a warning, not an error, and since you seem to be running your code with different input files than the ones in the OP (they don't have 14 lines), and I don't know what line 30 of your program does, I can only take a wild guess: Maybe $file2data{$columns1[0]} is empty, because the first columns of the files don't match up exactly. As stated, the example code doesn't really handle that case, but you could fairly easily modify it so it does, depending on what your input actually looks like. Without seeing a more representative sample of your input and your current program it's hard to say what the best solution is. See also the Basic debugging checklist.

You are wonderful. Thanks. I used my original input files and it works correctly. But i will try to learn Text::CSV. I hope i will find some useful examples about Text::CSV on internet.

my final code:

#!/usr/bin/perl 

use strict;
use warnings;
use diagnostics;


# Open file1 to read
open my $input_file1, '<', "rmsd1.xvg" or die qq{Failed to open "rmsd1
+.xvg" for writing: $!};

# Open file2 to read
open my $input_file2, '<', "rmsd2.xvg" or die qq{Failed to open "rmsd2
+.xvg" for writing: $!};

# Open new file to write
open my $out_file, '>', "rmsd.xvg" or die qq{Failed to open "rmsd.xvg"
+ for writing: $!};

# Copy data from one file to another.

my %file2data;
while(<$input_file2>)
{ 
 next if /(^\s*$)|(^#)|(^@)/;
 my @columns2 = split;
 $file2data{$columns2[0]} = $columns2[1];
}

while(<$input_file1>)
{ 
 next if /(^\s*$)|(^#)|(^@)/;
 my @columns1 = split;
 print $out_file join("\t", $columns1[0],$columns1[1], $file2data{$col
+umns1[0]}), "\n";
}
close($input_file2);
close($input_file1);
close($out_file);
[download]