Cascalog - std-dev using defaggregateop

gistfile1.clj

(defaggregateop variance
  "An implementation of the Welford/Knuth online algorithm:
   http://en.wikipedia.org/wiki/Algorithms_for_calculating_variance#On-line_algorithm"
  ([] [0 0 0])
  ([[m2 mean n] val]
    (let [n (inc n)
          delta (- val mean)
          mean (+ mean (div delta n))
          m2 (+ m2 (* delta delta))]
      [m2 mean n]))
  ([[m2 _ n]] [(div m2 (max 1 (- n 1)))]))

(defn sqrt [x] (Math/sqrt x))

;; Example of using it:
(?<- (stdout) [?key ?std-dev] (source ?key ?val) (variance ?val :> ?var) (sqrt ?var :> ?std-dev))

Prefer this version here: https://github.com/nathanmarz/cascalog-contrib/blob/master/cascalog.math/src/cascalog/math/stats.clj