sample-lines

#!/usr/bin/env perl
use v5.14;

# select a random sample of lines from a file
my $n = 1;
if ($ARGV[0] =~ /^-n(.*)/) {
  $n = $1;
  shift @ARGV;
  if (!$n) {
    $n = $ARGV[0];
    shift @ARGV;
  }
}

my @sample;
while (<>) {
  if (@sample < $n) {
    push @sample, $_;
  } else {
    my $j = int(rand $.);
    $sample[$j] = $_ if $j < $n;
  }
}
print @sample;

__END__
=head1 NAME

    sample-lines - Retrieve a random subset of input lines 

=head1 SYNOPSIS

    sample-lines [-n count] [file...]

=head1 OPTIONS

    -n	count   Specify the number of lines to return (default is 1 line)

=head1 DESCRIPTION

   Given any stream of lines of text, sample-lines will return a randomly-selected
   subset of any size.  If the requested size is more than the number of lines actually
   supplied, all input lines are returned; otherwise, each input line has an equal 
   probability of being included in the output sample.

=head1 AUTHOR

   Written by Mark J. Reed <mark.reed@turner.com>