timeseries/top_n

#!/usr/bin/perl -w 
use strict;

my $n = shift;
my $f = shift;
my $v = shift;

my @data;
my %s;
while (<>) {
    chomp;
    my @a = split(/\t/);
    push @data, [@a];
    $s{$a[$f]} += $a[$v];
}

my @top_n = (sort { $s{$b} <=> $s{$a} } keys %s )[ 0 .. $n - 1 ];
my %top_n = map { $_ => 1 } @top_n;

my %index;

for my $i (0 .. $#data) {
    unless ($top_n{$data[$i][$f]}) {
	my $val = $data[$i][$v];
	$data[$i][$v] = 0; 
	$data[$i][$f] = 'OTHER';
	my $k = join("\t", @{$data[$i]});
	if (defined($index{$k})) {
	    $data[$index{$k}][$v] += $val;
	    $data[$i][$v] = undef; 
	} else {
	    $data[$i][$v] = $val; 
	    $index{$k} = $i;
	}
    }
}

for (@data) {
    if (defined($_->[$v])) {
	print join("\t", @$_), "\n";
    }
}