Sums the rows and columns of a tab-delimited file

tsv_sum.pl

#!/usr/bin/env perl

#
# Reads a tab-delimited file with column and row names
# sums the rows and columns and outputs a tab-delimited
# file with an extra row and column for the sums
#

use strict;
use warnings;

my $usage = "Usage: $0 <infile.tsv>\n";
my $infile = shift or die $usage;

my $data = [];
my $header = '';
my $i = 1;

open(IN,'<',$infile) || die "Could not open $infile: $!\n";
while(<IN>){
   chomp;

   # expects header on the first line
   if ($. == 1){
      $header .= "$_\ttotal";
      next;
   }

   next if /^#/;
   next if /^$/;
   next if /^\s+/;

   my @s = split(/\t/);
   for (my $j = 0; $j < scalar(@s); ++$j){
      $data->[$i]->[$j] = $s[$j]
   }
   ++$i;
}
close(IN);

# variables for storing the column sums
my @col_sum = ();
my $last_col_sum = 0;

# print the header
print "$header\n";

# expects header, skipping first row
for (my $r=1; $r < scalar(@$data); ++$r){
   my $row_sum = 0;
   my $row = "$data->[$r]->[0]\t";
   # expects row names, skipping first column
   for (my $c=1; $c < scalar(@{$data->[$r]}); ++$c){
      $row .= "$data->[$r]->[$c]\t";
      $row_sum += $data->[$r]->[$c];

      # column sum
      $col_sum[$c] += $data->[$r]->[$c];
   }
   print "$row$row_sum\n";
   $last_col_sum += $row_sum;
}

my $col_sum_row = "total\t";
for (my $c = 1; $c < scalar(@col_sum); ++$c){
   $col_sum_row .= "$col_sum[$c]\t";
}

print "$col_sum_row$last_col_sum\n";

exit(0);