Hadoop Streaming with Ruby

map.rb

#!/usr/bin/env ruby

ARGF.each do |line|
  line.chomp!
  words = line.split(' ')
  printf("%s,1\n", words[8])
end

reduce.rb

#!/usr/bin/env ruby

counter = Hash.new
ARGF.each do |line|
  line.chomp!
  words = line.split(',')
  word = words[0]
  count = words[1].to_i
  if !counter.key?(word)
    counter[word] = 1
  else
    counter[word]+=count
  end
end

counter.each do |k, v|
  printf("%s:%d\n", k, v)
end