Good Ruby vs. Bad Ruby

good_vs_bad.rb

# Here are two Ruby programs that do exactly the same thing. One is written to be readable and maintainable, and the other is not.

#
# BAD RUBY
#

d = Hash.new([])
File.foreach('scale_test_timing_log.csv') { |line| d[line.split(',')[2]] = d[line.split(',')[2]] << line.chomp.split(',')[3].to_f }
d.keys.sort.each { |act| File.open('a.csv', 'a') { |a| a.puts("#{act},#{((d[act].sort[((d[act].length - 1) / 2.0).floor] + d[act].sort[((d[act].length - 1) / 2.0).ceil]) / 2.0).round(1)},#{d[act].last},#{d[act].select { |elapsed_secs| elapsed_secs >= 4 }.length}") } }


#
# GOOD RUBY
#

ANALYSIS_CSV                = 'scale_test_timing_analysis.csv'
DATA_CSV                    = 'scale_test_timing_log.csv'
MAX_ACCEPTABLE_ELAPSED_SECS = 4.0

data = Hash.new(Array.new)

# read in raw data
File.foreach(DATA_CSV) do |line|
  line_parts   = line.chomp.split(',')
  action       = line_parts[2]
  elapsed_secs = line_parts[3].to_f
  data[action] = data[action] << elapsed_secs
end

# prep the file that contains our analyzed results
File.write(ANALYSIS_CSV, "action,median elapsed secs,max elapsed secs,instances > #{MAX_ACCEPTABLE_ELAPSED_SECS} secs\n")

# analyze data and dump it to a file
actions = data.keys.sort
actions.each do |action|
  elapsed_time_in_secs = data[action].sort

  # find max elapsed time
  max_elapsed_secs     = elapsed_time_in_secs.last

  # find median elapsed time
  median_idx           = (elapsed_time_in_secs.length - 1) / 2.0
  median_elapsed_secs  = ((elapsed_time_in_secs[median_idx.floor] + elapsed_time_in_secs[median_idx.ceil]) / 2.0).round(1)

  # find number of instances that ran too slowly
  slow_instances       = elapsed_time_in_secs.select { |elapsed_secs| elapsed_secs >= MAX_ACCEPTABLE_ELAPSED_SECS }

  # write analyzed data to a CSV file
  File.open(ANALYSIS_CSV, 'a') do |analysis_csv|
    analysis_csv.puts("#{action},#{median_elapsed_secs},#{max_elapsed_secs},#{slow_instances.length}")
  end
end