Compares speed of two different methods of inserting items in an array.

item_insert_benchmark.rb

require 'benchmark'


def insert_then_sort(results, missing_items)
  missing_items.each do |id|
    results << {id: id, val: nil}
  end
  results.sort_by!{|item| item[:id] }
  results
end

def insert_using_cursors(results, missing_items)
  mi_cursor = 0
  rs_cursor = 0
  complete_set = []
  while mi_cursor < missing_items.count
    value_to_insert = missing_items[mi_cursor]
    mi_cursor += 1
    while rs_cursor < results.count && results[rs_cursor][:id] < value_to_insert
      complete_set << results[rs_cursor]
      rs_cursor += 1
    end
    complete_set << {id: value_to_insert, val: nil}
  end
  while rs_cursor < results.count ; complete_set << results[rs_cursor] ; rs_cursor += 1 ; end
  complete_set
end

results = []
missing_items = []
100_000.times do |n|
  if [true, false].sample
    results << {id: n, val: "fortitude"}
  else
    missing_items << n
  end
end

res1 = nil
res2 = nil
puts "INSERT THEN SORT"
puts Benchmark.measure { res1 = insert_then_sort(results.dup, missing_items.dup) }
puts "INSERT USING CURSORS"
puts Benchmark.measure { res2 = insert_using_cursors(results.dup, missing_items.dup) }


# puts
# puts "==== 1"
# puts res1
# puts "==== 2"
# puts res2

unless res1 == res2
  puts "WARNING: The result sets were not identical!"
end

__END__

With 10,000 items
=================
INSERT THEN SORT
  0.150000   0.000000   0.150000 (  0.161145)
INSERT USING CURSORS
  0.040000   0.010000   0.050000 (  0.051906)

With 1,000,000 items
=================
INSERT THEN SORT
  2.430000   0.050000   2.480000 (  2.577645)
INSERT USING CURSORS
  1.040000   0.040000   1.080000 (  1.076837)