aggregator.rb

# input:
#
# foo,1,1
# foo,1,2
# foo,1,3
# foo,1,5
# bar,2,1
# bar,1,7
# bar,0,5
#
#
# output:
# 
# foo,1.0,2.5
# bar,1.0,5.0

a1 = Aggregator.new do |agg|
  agg.group_by 1
  agg.all_columns do |data|
    data.median
  end
end

# other ideas:

# this would take a file with 5 columns, group on the first two,
# and then output 4 data columns:
#   * the median of column 3
#   * the mean of column 4
#   * the median of column 5
#   * the mean of column 5
a2 = Aggregator.new do |agg|
  agg.group_by 1, 2
  agg.column 3 do |data|
    data.median
  end
  agg.column 4 do |data|
    data.mean
  end
  agg.column 5 do |data|
    data.median
  end
  agg.column 5 do |data|
    data.mean
  end
end

# this illustrates optional column naming
a3 = Aggregator.new do |agg|
  agg.columns :foo, :bar, :baz
  agg.group_by :foo
  agg.column :bar, :baz do |data|
    data.mean
  end
end

a4 = Aggregator.new do |agg|
  agg.sum(column(1))
  agg.column(1, &sum)
  agg.out column(1).sum
end

class AggClass < Aggregate
  group_by :foo
  
  output(:foo).sum
  output(:foo).avg
  output(:foo).median
  
  output(:foo) do |data| # loads all data
    data.first * 2
  end
  
  output(:foo).inject(0) do |acc, rec| # streams data
    acc += rec % 10
  end
end

a5 = AggClass.new
a5.aggregate($stdin)

agg_class = Class.new(Aggregate) do
  group_by :foo
  
  output(:foo).sum
  output(:foo).avg
  output(:foo).median
  
  output(:foo) do |data| # loads all data
    data.first * 2
  end
  
  output(:foo).inject(0) do |acc, rec| # streams data
    acc += rec % 10
  end
end

a6 = agg_class.new
a6.aggregate($stdin)