Skip to content

Latest commit

 

History

History
517 lines (373 loc) · 11.2 KB

functions.md

File metadata and controls

517 lines (373 loc) · 11.2 KB
Raw

Functions

Creating Functions

(do
   (defn add [x y] (+ x y))
   
   (def mul (fn [x y] (* x y)))
   
   (add 1 2)
   (mul 3 4)
   (let [f (fn [x y] (- x y))] 
      (f 5 3)))

Variadic Functions

A variadic function is a function of indefinite arity, accepting a variable number of arguments. A variadic function can have any number of fixed arguments.

(do
   ;; variadic sum with a single fixed arg
   (defn sum [x & xs]
      (reduce + x xs))

   (sum 1)
   (sum 1 2)
   (sum 1 2 3 4 5 6 7 8 9 10))

Multi-Arity Functions

(do
   (defn arity
      ([] (println "arity 0"))
      ([a] (println "arity 1"))
      ([a b] (println "arity 2"))
      ([a b c] (println "arity 3"))
      ([a b c & z] (println "arity 3+")))
      
   (arity 1 2))

Passing Values by Name to Functions

Values can be passed by name using a map literal.

The function foo expects the named arguments x and y. The :or clause specifies a default for y.

(do
   (defn foo [{:keys [x y] :or {y 10}}] 
      (list x y))
      
   (foo {:x 1 :y 2})  ;; => (1 2)
   (foo {:x 1}))      ;; => (1 10)

Anonymous Functions

(do
   (map (fn [x] (* 2 x)) (range 0 5))   ; => (0 2 4 6 8)
   (map #(* 2 %) (range 0 5))           ; => (0 2 4 6 8)
   (map #(* 2 %1) (range 0 5))          ; => (0 2 4 6 8)
   (let [f #(+ %1 %2 %3)] (f 1 2 3))    ; => 6
   ((fn [x] (* x 10)) 5))               ; => 50

Applying Functions

The apply function invokes a function with 0 or more fixed arguments, and draws the rest of the needed arguments from a list or a vector. The last argument must be a list or a vector.

(do
  (apply str '(1 2 3 4))    ;; same as (str 1 2 3 4)
  (apply str 1 '(2 3 4))    ;; same as (str 1 2 3 4)
  (apply str 1 2 '(3 4))    ;; same as (str 1 2 3 4)
  (apply str 1 2 3 '(4)))   ;; same as (str 1 2 3 4)

Locals and Closures

let

let binds symbols to values in a "lexical scope". A lexical scope creates a new context for names, nested inside the surrounding context. Names defined in a let take precedence over the names in the outer context.

(let [x 1
      y (* 2 3)]
  (+ x y))

This let expression creates two local bindings for x and y. The expression (+ x y) is in the lexical scope of the let and resolves x to 1 and y to 6. Outside the "lexical scope" (let expression), x and y will not be accessible.

defn

defn binds its function arguments in a new "lexical scope" as well. Unlike let the local bindings for x and y get their values from the caller.

(do
  (defn sum [x y]
     (+ x y))
   
  (sum 1 2))

Closures

The fn special form creates a "closure". It "closes over" the surrounding lexical scope and captures their values beyond the lexical scope.

A closure is a function that remembers the environment at which it was created.

(do
  (defn pow [n]
    (fn [x] (apply * (repeat n x))))  ; closes over n

  ;; n is provided here as 2 and 3, then n goes out of scope
  (def square (pow 2))
  (def cubic (pow 3))
  
  ;; n value still available because square and cubic are closures
  (square 4) ; => 16   effectively as (apply * (repeat 2 4)) 
  (cubic 4)) ; => 64   effectively as (apply * (repeat 3 4))

Even global functions can remember the context they have been created:

(do
  (let [x 100]
    (defn test [] (str "x: " x)))
    
  (test))  ; => "x: 100"

Functions with preconditions

(do
   (defn sum [x y] 
      { :pre [(number? x) (number? y)] } 
      (+ x y)))

Argument type hints

Venice supports function argument type hints through argument metadata. Type hints are available with Venice 1.11.x.

(do
  (defn sum [^:long x ^:long y] (+ x y))
  (sum 1 2))
(do
  (defn sum [^:number x ^:number y] (+ x y))
  
  (sum 1 2)
  (sum 1.0 2)
  (sum 1.1M 2.6M))
(do
  (ns foo)
  (deftype :complex [real      :long
                     imaginary :long])
                     
  (defn sum [^:foo/complex x ^:foo/complex y] 
    (complex. (+ (:real x) (:imaginary x)) 
              (+ (:real y) (:imaginary y))))
     
  (sum (complex. 1 2) (complex. 5 8)))

Type hints with multi-arity functions:

(do
   (defn foo
      ([] 0)
      ([^:long x] x)
      ([^:long x ^:long y] (+ x y))
      ([^:long x ^:long y & xs] (apply + x y xs)))
   (foo )
   (foo 1)
   (foo 1 2)
   (foo 1 2 3 4 5))

Type hints with sequential destructuring:

(do
   (defn foo [[^:long x ^:long y]] (+ x y))
   (foo [1 2]))

Type hints with associative destructuring:

(do
   (defn foo [{:keys [^:long x ^:long y]}] (+ x y))
   (foo {:x 1 :y 2}))

For datatypes of the core namespace the namespace can be omitted.

;; these two function definitions are equivalent
(defn sum [^:long x ^:long y] (+ x y)))
(defn sum [^:core/long x ^:core/long y] (+ x y)))

Collections and Keywords as functions

Vectors, maps, sets, and keywords are functions too.

Vectors

(do
   ([1 2 3] 1) ; -> 2
   
   ;; with default
   ([1 2 3] 5 10)) ; -> 10

Maps

(do
   ;; instead of (get {:a 1 :b 2} :b)
   ({:a 1 :b 2} :b)  ; -> 2
   ({:a 1 :b 2} :c)  ; -> nil
   
   ;; with default
   ({:a 1 :b 2} :c 9)  ; -> 9

   ;; accessing nested maps
   (let [m {:a {:b {:c 3}}}]
      (-> m :a :b :c)))  ; -> 3

Sets

(do
   (#{:a :b} :b)  ; -> :b
   (#{:a :b} :c)  ; -> nil
   
   ;; with default
   (#{:a :b} :c 9))  ; -> 9

Keywords

(do
   (:b {:a 1 :b 2})  ; -> 2
   (:c {:a 1 :b 2})  ; -> nil
   (:b #{:a :b})  ; -> :b
   (:c #{:a :b})  ; -> nil
 
   ;; with defaults
   (:c {:a 1 :b 2} 9)  ; -> 9
   (:c #{:a :b} 9))    ; -> 9

Function resolved from a string

(let [add (resolve (symbol "+"))]
   (add 2 5))

Partial Functions

In computer science, partial application (or partial function application) refers to the process of fixing a number of arguments to a function, producing another function of smaller arity.

(do
  (defn add [x y] (+ x y))
  
  (def add2 (partial add 2))
  (def add3 (partial add 3))
  
  (add2 4)   ;; => 6
  (add3 4))  ;; => 7

Using functions with fewer arguments than they can normally take

If we did need to call add (adding two numbers) with fewer than the required arguments, for example if we are mapping add over a vector, then we can use partial to help us call the add function with the right number of arguments:

(do
  (defn add [x y] (+ x y))
  (map (partial add 2) [1 2 3 4]))  ;; => (3 4 5 6)

In this case the partial function prevents us from writing an explicit anonymous function like #(+ 2 %) in (map #(+ 2 %) [1 2 3 4])

(map (partial + 2) [1 2 3 4])  ;; => (3 4 5 6)

Using functions with more arguments than they can normally take

The reduce function can only work on a single collection as an argument (or a value and a collection), so an error occurs if you wish to reduce over multiple collections.

(reduce + [1 2 3 4])  ;; => 10

This returns an error due to invalid arguments:

(reduce + [1 2 3 4] [5 6 7 8])  ;; error

However, by using partial we can take one collection at once and return the result of reduce on each of those collections:

(map (partial reduce +) [[1 2 3 4] [5 6 7 8]])  ;; => (10 26)

Function Composition

Creates a new function composed of one or more functions. The composed functions are executed from right to left.

(comp not zero?) is equivalent to (fn [x] (not (zero? x))

(filter (comp not zero?) [0 1 0 2 0 3 0 4])  ;; => [1 2 3 4]
(do
  (def person
    {:name "Peter Meier"
     :address {:street "Lindenstrasse 45"
               :city "Bern" 
               :zip 3000}})
  ((comp :street :address) person) ;; => "Lindenstrasse 45"
  ((comp :private :email) person)) ;; => nil
(do
  (def xform
    (comp 
      (partial take 4)
      (partial map #(+ 2 %))
      (partial filter odd?)))
    
  (xform (range 0 10)))  ;; => (3 5 7 9)

Function threading macros

Thread first ->

Taking an initial value as its first argument, -> threads it through one or more expressions. Starting with the second form, the macro inserts the first value as its first argument and repeats inserting the result of the form to the first argument of the next form.

(do
  (defn bigint [x] (. :java.math.BigInteger :new x))
  (-> (bigint "1000")
      (. :multiply (bigint "600"))
      (. :add (bigint "300"))))  ;; => 600300
(do
  (def person
    {:name "Peter Meier"
     :address {:street "Lindenstrasse 45"
               :city "Bern" 
               :zip 3000}})
  (-> person :address :street) ;; => "Lindenstrasse 45"
  (-> person :email :private)) ;; => nil

Thread last ->>

Taking an initial value as its first argument, ->> threads it through one or more expressions. Starting with the second form, the macro inserts the first value as its last argument and repeats inserting the result of the form to the last argument of the next form.

(->> (range 0 8)
     (filter odd?)
     (map #(+ 2 %)))  ;; => (3 5 7 9)

Thread any as->, -<>

; allows to use arbitrary positioning of the argument
(as-> (range 0 8) v
      (filter odd? v)
      (reduce + v)
      (* v 2))  ;; => 32
; the chosen threading symbol may be used multiple times in a form
; thus allowing the use of complex forms like if expressions
(as-> {:a 1 :b 2} m
      (update m :a #(+ % 10))
      (if true
        (update m :b #(+ % 10))
         m))  ;; => {:a 11 :b 12}
; allows to use arbitrary positioning of the argument using the placeholder '<>'
; note: the threading symbol <> may only be used once in a form
(-<> (range 0 8)
     (filter odd? <>)
     (reduce + <>)
     (* <> 2))  ;; => 32

Functions calling each other

Venice supports functions calling each other without needing to declare them or bind them with letfn as required with Clojure.

Nevertheless alternately calling functions can cause stack overflows if the recursion is too deep. Use the trampoline function or convert to self-recursion to overcome the stack overflow problem.

(do
  (let [print-number (fn [n]
                         (println n)
                         (decrease n))
        decrease (fn [n]
                     (when (> n 0)
                       (print-number (dec n))))]
    (decrease 10)))