Re: Bucketing,Slicing and Reporting data across multiple dimensions

[blue_cowdawg]

This may be an oversimplification, but here is one idea:

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

open FIN,"<dataset.dat" or die "dataset.dat:$!";
my $buckets={};

while(my $line=<FIN>){
    chomp $line;
    my @f=split(",",$line);
    unless (exists($buckets->{$f[1]}){
        $buckets->{$f[1]} = { count => 0, where =>[ ] };
    }
    if ( $f[4] > 0 ) { 
         $buckets->{$f[1]}->{count}++;
         push @{$buckets->{$f[1]}},$f[0];
    }
}
close FIN;
foreach my $key (sort keys %$buckets){
      next unless $buckets->{$key}->{count} > 0 ;
      printf "%s\t%d\t%s\n",
                $key,$buckets->{$key}->{count},
                join(",",sort @{$buckets->{$key}->{where}});
}

#
####################################################### done

Caveat: not tested.. do not use for the control of a nuclear missile launch system.

[Voronich]

Thanks very much! This looks remarkably close to what I'm looking for. Debugging now. (It's complaining about that "push" not referencing an array ref.)

[blue_cowdawg]

push @{$buckets->{$f[1]}},$f[0];

Doh!!! It should read:

push @{$buckets->{$f[1]}->{where}},$f[0];

[Voronich]

Ah. Domo o/

[Voronich]

The only other modification I made was:

    printf "%s\t%d\n",
    $key,$buckets->{$key}->{count};

since I don't actually need the articulated list of IDs unless they're called for. The resulting list is really quite interesting. Almost all impact occurs across EVERY ID. But several only have impact on 3 or 4. Quite telling. Thanks again o/