Answering a question in SO, I stumbled into this problem:
(def x [7 4 8 9 10 54 55 2 23 30 12 5])
(defn insert-x 
  ([sorted-coll x] 
   (insert-x sorted-coll x 
     (if (= (type sorted-coll) clojure.lang.PersistentVector) [] '())))
  ([sorted-coll x acc]
  (let [is-vector  (= (type sorted-coll) clojure.lang.PersistentVector)
        format-it  #(into (if is-vector [] '()) %)
        compare   (if is-vector < >)]
    (cond 
      (empty? sorted-coll) (format-it (cons x acc))
      (compare (peek sorted-coll) x) 
      (format-it (concat 
                   ((if is-vector identity reverse) sorted-coll) 
                   (conj acc x)))
      :else (recur (pop sorted-coll) x (cons (peek sorted-coll) acc))))))
(defn bubble-sort [coll]
  "Insert x into a sorted collection"
  (reduce insert-x [] coll))
(bubble-sort x)
;; => [2 4 5 7 8 9 10 12 23 30 54 55]
The code does what it should.
However, insert-x is not so elegant.
How to write insert-x in a way that it is valid for all collections?
So that it is simpler/more elegant?
vectors should return vectors, lists should return lists etc.
The best general purpose or 'primary' implementations are likely ArrayList , LinkedHashMap , and LinkedHashSet . Their overall performance is better, and you should use them unless you need a special feature provided by another implementation. That special feature is usually ordering or sorting.
i guess you're overthinking it.
you have two tasks:
first of all, i would rewrite the insert-x like this for example:
(defn insert-x [sorted-coll x]
  (let [[l r] (split-with #(<= % x) sorted-coll)]
    `(~@l ~x ~@r)))
notice, it does more or less the same that your variant does: taking values until the desired positions, and then concatenating left and right parts with x between them. notice also, it always produces properly sorted list, independent from the input type.
user> (insert-x [1 3 5 7 9] 10)
;;=> (1 3 5 7 9 10)
user> (insert-x [1 3 5 7 9] 0)
;;=> (0 1 3 5 7 9)
user> (insert-x [1 3 5 7 9] 4)
;;=> (1 3 4 5 7 9)
so, the next thing you need, is just to reduce input and return the properly typed result:
(defn my-sort [coll]
  (let [sorted (reduce insert-x () coll)]
    (if (vector? coll)
      (vec sorted)
      sorted)))
user> (my-sort '(0 3 1 4 2 5 10 7))
;;=> (0 1 2 3 4 5 7 10)
user> (my-sort [0 3 1 4 2 5 10 7])
;;=> [0 1 2 3 4 5 7 10]
user> (my-sort ())
;;=> ()
user> (my-sort [])
;;=> []
                        What you need is to have insert-x working on regular list (ie. '() or nil) and rewriting bubble-sort to create the same type as input using empty function.
empty returns empty collection of the same type:
(class (empty [1 2]))
;; => clojure.lang.PersistentVector
(class (empty #{1 2}))
;; => clojure.lang.PersistentHashSet
(class (empty '(1 2)))
;; => clojure.lang.PersistentList$EmptyList
Your bubble-sort can look this way:
(defn bubble-sort [coll]
  "Insert x into a sorted collection"
  (into (empty coll) (reduce insert-x nil coll)))
This way you can get rid of all type checks inside insert-x
Thank you both!
I took both of your answers and generated this solution:
(defn insert-x [sq x]
  (apply concat (interleave (split-with #(<= % x) sq) [[x] []])))
(defn bubble-sort [sq]
  ((if (vector? sq) vec identity)
   (reduce insert-x () sq)))
Test:
user=> (bubble-sort '(0 2 8 3 9 7))
(0 2 3 7 8 9)
user=> (bubble-sort [0 2 8 3 9 7])
[0 2 3 7 8 9]
Or probably even better readable:
(defn insert [q x]
  (let [[l r] (split-with #(< % x) sq)] 
    (concat l [x] r)))
(defn bubble-sort [sq]
  ((if (vector? sq) vec identity) 
   (reduce insert () sq)))
Or:
(defn insert [sq x]
  (let [[l r] (split-with #(< % x) sq)] 
    `(~@l ~x ~@r)))
(defn intoo [x sq] ;; direction neutral `into`
  (into x (into x sq))) 
(defn bubble-sort [sq]
  (intoo (empty sq) (reduce insert () sq)))
intoo is needed because of
(into () '(1 2 3)) ;;=> '(3 2 1)
(into () [1 2 3])  ;;=> '(3 2 1)
(into [] '(1 2 3)) ;;=> [1 2 3]
(into [] [1 2 3])  ;;=> [1 2 3]
# `into` uses `conj` underneath!
Apply two times successively into fixes this:
(into () (into () '(1 2 3))) ;; '(1 2 3)
(into () (into () [1 2 3]))  ;; '(1 2 3)
(into [] (into [] '(1 2 3))) ;; [1 2 3]
(into [] (into [] [1 2 3]))  ;; [1 2 3]
Thus intoo is kind of a direction neutral version of into.
Actually, the most performance efficient version of intoo might be:
(defmacro intoo [x sq]
  (if (vector? x)
      `(into ~x ~sq)
      `(reverse (into ~x ~sq))))
;; user=> (intoo () '(1 2 3))
;; (1 2 3)
;; user=> (intoo () [1 2 3])
;; (1 2 3)
;; user=> (intoo [] [1 2 3])
;; [1 2 3]
;; user=> (intoo [] '(1 2 3))
;; [1 2 3]
;; or as a function:
(defn intoo [x & args]
  (if (vector? x)
      (apply into x args)
      (reverse (apply into x args))))
                        If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With