A simple Ruby script illustrating how to implement Redis-style ZRANK on the result of a Keen IO group_by query

keen_io_zrank.rb

# Example Keen IO API response for a count, grouped by an 'entry' field
keen_result = [
  {
    "entry" => "Apple",
    "result" => 15
  },
  {
    "entry" => "Orange",
    "result" => 10
  },
  {
    "entry" => "Grape",
    "result" => 5
  },
  {
    "entry" => "Pear",
    "result" => 7
  }
]

# flatten the API response into a single hash
# e.g. {"Apple"=>15, "Orange"=>10, "Grape"=>5, "Pear"=>7}
flat_hash = keen_result.inject({}) { |acc, entry|
  acc[entry["entry"]] = entry["result"]; acc
}

# create a 'set', sorted by the entry count
# e.g. [["Grape", 5], ["Pear", 7], ["Orange", 10], ["Apple", 15]]
sorted_set = flat_hash.sort_by { |entry, count| count }

# given an sorted array of pairs, find the index of the pair
# whose value matches the 'entry' param
def zrank(entry, set)
  set.each_with_index { |pair, index|
    return index if entry == pair[0]
  }
end

puts "Apple:  #{zrank("Apple",  sorted_set)}" #=> 3
puts "Orange: #{zrank("Orange", sorted_set)}" #=> 2
puts "Grape:  #{zrank("Grape",  sorted_set)}" #=> 0
puts "Pear:   #{zrank("Pear",   sorted_set)}" #=> 1

The Keen IO API doesn't have a formal ZRANK operation but it's fairly easy to implement using the response of a group by query.

Thanks :)