Book Image

Clojure Reactive Programming

Book Image

Clojure Reactive Programming

Overview of this book

Table of Contents (19 chapters)
Clojure Reactive Programming
Credits
About the Author
Acknowledgments
About the Reviewers
www.PacktPub.com
Preface
Bibliography
Index

Gathering stats about ages


Now that we can safely calculate the average age of a number of pirates, it might be interesting to take this further and calculate the median and standard deviation of the pirates' ages, in addition to their average age.

We already have a function to calculate the average, so let's just create the ones to calculate the median and the standard deviation of a list of numbers:

(defn median [& ns]
  (let [ns (sort ns)
        cnt (count ns)
        mid (bit-shift-right cnt 1)]
    (if (odd? cnt)
      (nth ns mid)
      (/ (+ (nth ns mid) (nth ns (dec mid))) 2))))

(defn std-dev [& samples]
  (let [n (count samples)
	mean (/ (reduce + samples) n)
	intermediate (map #(Math/pow (- %1 mean) 2) samples)]
    (Math/sqrt
     (/ (reduce + intermediate) n))))

With these functions in place, we can write the code that will gather all the stats for us:

  (let  [a       (some-> (pirate-by-name "Jack Sparrow")    age)
         b       (some-> (pirate-by-name "Blackbeard")      age)
         c       (some-> (pirate-by-name "Hector Barbossa") age)
         avg     (avg a b c)
         median  (median a b c)
         std-dev (std-dev a b c)]
    {:avg avg
     :median median
     :std-dev std-dev})

  ;; {:avg 56.666668,
  ;;  :median 60,
  ;;  :std-dev 12.472191289246473}

This implementation is fairly straightforward. We first retrieve all ages we're interested in and bind them to the locals a, b, and c. We then reuse the values when calculating the remaining stats. We finally gather all results in a map for easy access.

By now the reader will probably know where we're headed: what if any of those values is nil?

  (let  [a       (some-> (pirate-by-name "Jack Sparrow")    age)
         b       (some-> (pirate-by-name "Davy Jones")      age)
         c       (some-> (pirate-by-name "Hector Barbossa") age)
         avg     (avg a b c)
         median  (median a b c)
         std-dev (std-dev a b c)]
    {:avg avg
     :median median
     :std-dev std-dev})
  ;; NullPointerException   clojure.lang.Numbers.ops (Numbers.java:961)

The second binding, b, returns nil, as we don't have any information about Davy Jones. As such, it causes the calculations to fail. Like before, we can change our implementation to protect us from such failures:

  (let  [a       (some-> (pirate-by-name "Jack Sparrow")    age)
         b       (some-> (pirate-by-name "Davy Jones")      age)
         c       (some-> (pirate-by-name "Hector Barbossa") age)
         avg     (when (and a b c) (avg a b c))
         median  (when (and a b c) (median a b c))
         std-dev (when (and a b c) (std-dev a b c))]
    (when (and a b c)
      {:avg avg
       :median median
       :std-dev std-dev}))
  ;; nil

This time it's even worse than when we only had to calculate the average; the code is checking for nil values in four extra spots: before calling the three stats functions and just before gathering the stats into the result map.

Can we do better?