How to identify nth lines of n files in while<> - perl

Question

I have a code which adds all vectors in all files.
There can be any number of input files. For example first input file is:
0.55 0 0.3335 1.2
0.212 0 2.2025 1
and the second one is:
0.25 0 0.3333 1.0
0.1235 0 0.2454 1
What I get is the sum of all vectors, thus in result i get one vector
which is:
1.13550 0 3.1147 4.2
But I'm trying to sum the first vector of the first file with the first vector of the second file and so on. In result according to this example I should get 2 vectors.
For now I have this:
use strict;
use warnings;
if ($ARGV[0] ne "vector1.dat"){
die ("vector1.dat is necessary as first argument");
}
my #sum = 0;
my $dim = 0;
while (<>) {
#Ignore blank lines, hashtags
#and lines starting with $
if ($_ =~ /#/ || $_ =~ /^$/ || $_ =~ /^\s$/){
next;
}
my #vectors = split(" ", $_);
my $vector_length = #vectors;
if ($dim eq 0) {
$dim = $vector_length;
}
else {
if ($dim ne $vector_length) {
die ("Vector dimensions do not match. : $!");
}
}
for (my $i = 0; $i <= $#vectors; $i++) {
$sum[$i] += $vectors[$i];
}
}
$" = "\t\t";
print "\n --- \n #sum \n";
What I need is just to find out how to identify each file's nth line
and to sum the column values of those lines while keeping in mind, that there can be n number of files.
I saw filehandling question over here with similar issue, however
I didn't find my answer there.
Just looking for some suggestions and guidance. Got stuck on this.

Answer 1

Open each file yourself and use the $. variable to know which line you are on (or count the files yourself). Here's the basic structure:
foreach my $file ( #files ) {
open my $fh, '<', $file or die ...;
while( <$fh> ) {
chomp;
$sum[ $. ] = ...; # $. is the line number
}
}
If you don't like $., you can use its longer name. You have to turn on English (which comes with Perl):
use English;
## use English qw( -no_match_vars ); # for v5.16 and earlier
foreach my $file ( #files ) {
open my $fh, '<', $file or die ...;
while( <$fh> ) {
chomp;
$sum[ $INPUT_LINE_NUMBER ] = ...;
}
}
Or, you can count yourself, which might be handy if the vectors in the files don't line up by strict line number (perhaps because of comments or some other formatting oddity):
foreach my $file ( #files ) {
open my $fh, '<', $file or die ...;
my $line = -1;
while( <$fh> ) {
$line++;
chomp;
$sum[ $line ] = ...;
}
}
The harder way is the answer bart gives which inspects eof at the end of every line to see if the magical ARGV handle is looking at a new file, and resetting $. if it is. It's an interesting trick but hardly anyone is going to understand what it's doing (or even notice it).
For the other part of the problem, I think you're doing the vector sum wrong, or using confusing variable names. A line is a vector, and the numbers in the lines are a component. A two dimensional array will work. The first index is the line number and the second in the component index:
while( <$fh> ) {
chomp;
... skip unwanted lines
my #components = split;
... various dimension checks
foreach my $i ( 0 .. $#components ) {
$sum[ $. ][ $i ] += $components[ $i ];
}
}
The Data::Dumper module is handy for complex data structures. You can also see the perldsc (Perl Data Structures Cookbook) documentation. The $. variable is found in perlvar.

Answer 2

$. is the line number of the most recently read file handle. close(ARGV) if eof; can be used to reset the file number between files (as documented in eof). (Note: eof() is different than eof.) So you now have line numbers.
The second problem you have is that you are adding vector components ($vectors[$i]) to a vectors ($sum[$i]). You need to add vector components to vectors components. Start by using more appropriate variable names.
This is what we get:
my #sum_vectors;
while (<>) {
s/#.*//; # Remove comments.
next if /^\s*$/; # Ignore blank lines.
my #vector = split;
if ($sum_vectors[$.] && #{ $sum_vectors[$.] } != #vector) {
die("$ARGV:$.: Vector dimensions do not match\n");
}
for my $i (0..$#vector) {
$sum_vectors[$.][$i] += $vector[$i];
}
} continue {
close(ARGV) if eof; # Reset line numbers for each file.
}
Two other errors fixed:
$! did not contain anything meaningful when you used it.
You ignored lines that contain comments, even if they contained valid data too.

Answer 3

Try this:
#!/usr/bin/perl
use strict;
use warnings;
if ($ARGV[0] ne "vector1.dat"){
die ("vector1.dat is necessary as first argument");
}
my %sum;
my $dim = 0;
my $vector_length;
my $line_number;
while (<>) {
#Ignore blank lines, hashtags
#and lines starting with $
if ($_ =~ /#/ || $_ =~ /^$/ || $_ =~ /^\s$/){
next;
}
my #vectors = split(" ", $_);
$vector_length = #vectors;
if ($dim eq 0) {
$dim = $vector_length;
}
else {
if ($dim ne $vector_length) {
die ("Vector dimensions do not match. : $!");
}
}
for (my $i = 0; $i <= $#vectors; $i++) {
$sum{$.}{$i} += $vectors[$i];
}
$line_number = $.;
$. = 0 if eof;
}
$" = "\t\t";
for (my $line=1; $line<=$line_number; $line++)
{
print $line;
for (my $vector=0; $vector<$vector_length; $vector++)
{
print " " . $sum{$line}{$vector};
}
print "\n";
}