colstat

#!/usr/bin/perl
use strict;
use Getopt::Std;
use FindBin;use lib $FindBin::Bin;

my $usage = q/Usage:
 colstat [-d <delim>] [-s <skip>] [-n] -f<columns> file..
 Reports sum, min, max and mean values for specific column(s) in the input.
 Options:
  -d  : columns are delimited by <delim> (regular expression format); 
        default is the tab character ("\t")
  -f  : provide a list of column numbers to report values for
  -s  : skip the initial <skip> lines in the input file
  -n  : ignore non-numeric or empty fields
/;
umask 0002;
getopts('nd:s:f:o:') || die($usage."\n");
my $outfile=$Getopt::Std::opt_o;
if ($outfile) {
  open(OUTF, '>'.$outfile) || die("Error creating output file $outfile\n");
  select(OUTF);
}
#my $cmd=shift(@ARGV) || die("$usage\n"); 
my $bN=2**63;
my $delim=$Getopt::Std::opt_d || "\t";

my $cols=$Getopt::Std::opt_f || die("${usage}Error: -f option is required!\n");
my $lskip=$Getopt::Std::opt_s;
#$lskip-- if $lskip>0;
my $ncheck=$Getopt::Std::opt_n;
my @f=($cols=~m/(\d+)/g);
my @out; #list of [colidx, count, sum, min, max]
foreach my $c (@f) {
  push(@out, [$c, 0, 0, $bN, -$bN]);
}

while (<>) {
  next if $. < $lskip;
  chomp;
  my @t=split(/$delim/);
  foreach my $d (@out) {
   my $v=@t[$$d[0]-1];
   next if $ncheck && $v!~m/[\-\d\.\+\-E]+/i;
   $$d[1]++;
   $$d[2]+=$v;
   $$d[3]=$v if $v<$$d[3];
   $$d[4]=$v if $v>$$d[4];
  }
}
#print results:
foreach my $d (@out) {
 my $avg=$$d[2]/$$d[1];
 print "Column $$d[0]: $$d[1] values\tsum=$$d[2]\tmin=$$d[3]\tmax=$$d[4]\tavg=$avg\n";
}

# --
if ($outfile) {
  select(STDOUT);
  close(OUTF);
}