2013-06-26 17:58:50 +00:00
|
|
|
---
|
|
|
|
language: clojure
|
2013-07-01 21:50:59 +00:00
|
|
|
filename: learnclojure.clj
|
2013-07-04 05:59:13 +00:00
|
|
|
contributors:
|
|
|
|
- ["Adam Bard", "http://adambard.com/"]
|
2013-06-26 17:58:50 +00:00
|
|
|
---
|
|
|
|
|
2013-07-01 11:58:34 +00:00
|
|
|
Clojure is a Lisp family language developed for the Java Virtual Machine. It has
|
2013-06-26 18:19:21 +00:00
|
|
|
a much stronger emphasis on pure [functional programming](https://en.wikipedia.org/wiki/Functional_programming) than
|
|
|
|
Common Lisp, but includes several [STM](https://en.wikipedia.org/wiki/Software_transactional_memory) utilities to handle
|
|
|
|
state as it comes up.
|
|
|
|
|
|
|
|
This combination allows it to handle concurrent processing very simply,
|
|
|
|
and often automatically.
|
2013-06-26 17:58:50 +00:00
|
|
|
|
2013-06-28 22:56:52 +00:00
|
|
|
(You need a version of Clojure 1.2 or newer)
|
|
|
|
|
|
|
|
|
2013-06-26 17:58:50 +00:00
|
|
|
```clojure
|
|
|
|
; Comments start with semicolons.
|
|
|
|
|
|
|
|
; Clojure is written in "forms", which are just
|
|
|
|
; lists of things inside parentheses, separated by whitespace.
|
|
|
|
;
|
2015-05-13 09:00:57 +00:00
|
|
|
; The clojure reader assumes that the first thing is a
|
2013-06-26 17:58:50 +00:00
|
|
|
; function or macro to call, and the rest are arguments.
|
2013-07-01 21:50:59 +00:00
|
|
|
|
|
|
|
; The first call in a file should be ns, to set the namespace
|
|
|
|
(ns learnclojure)
|
2013-06-26 17:58:50 +00:00
|
|
|
|
|
|
|
; More basic examples:
|
|
|
|
|
|
|
|
; str will create a string out of all its arguments
|
|
|
|
(str "Hello" " " "World") ; => "Hello World"
|
|
|
|
|
|
|
|
; Math is straightforward
|
|
|
|
(+ 1 1) ; => 2
|
|
|
|
(- 2 1) ; => 1
|
|
|
|
(* 1 2) ; => 2
|
|
|
|
(/ 2 1) ; => 2
|
|
|
|
|
|
|
|
; Equality is =
|
|
|
|
(= 1 1) ; => true
|
|
|
|
(= 2 1) ; => false
|
|
|
|
|
|
|
|
; You need not for logic, too
|
|
|
|
(not true) ; => false
|
|
|
|
|
|
|
|
; Nesting forms works as you expect
|
|
|
|
(+ 1 (- 3 2)) ; = 1 + (3 - 2) => 2
|
|
|
|
|
|
|
|
; Types
|
|
|
|
;;;;;;;;;;;;;
|
|
|
|
|
|
|
|
; Clojure uses Java's object types for booleans, strings and numbers.
|
|
|
|
; Use `class` to inspect them.
|
|
|
|
(class 1) ; Integer literals are java.lang.Long by default
|
|
|
|
(class 1.); Float literals are java.lang.Double
|
|
|
|
(class ""); Strings always double-quoted, and are java.lang.String
|
|
|
|
(class false) ; Booleans are java.lang.Boolean
|
|
|
|
(class nil); The "null" value is called nil
|
|
|
|
|
2013-06-29 22:37:42 +00:00
|
|
|
; If you want to create a literal list of data, use ' to stop it from
|
|
|
|
; being evaluated
|
2013-06-26 17:58:50 +00:00
|
|
|
'(+ 1 2) ; => (+ 1 2)
|
2013-10-03 21:31:56 +00:00
|
|
|
; (shorthand for (quote (+ 1 2)))
|
2013-06-26 17:58:50 +00:00
|
|
|
|
2013-06-29 22:37:42 +00:00
|
|
|
; You can eval a quoted list
|
2013-06-26 17:58:50 +00:00
|
|
|
(eval '(+ 1 2)) ; => 3
|
|
|
|
|
|
|
|
; Collections & Sequences
|
|
|
|
;;;;;;;;;;;;;;;;;;;
|
|
|
|
|
2013-07-01 21:50:59 +00:00
|
|
|
; Lists are linked-list data structures, while Vectors are array-backed.
|
2013-06-26 17:58:50 +00:00
|
|
|
; Vectors and Lists are java classes too!
|
|
|
|
(class [1 2 3]); => clojure.lang.PersistentVector
|
|
|
|
(class '(1 2 3)); => clojure.lang.PersistentList
|
|
|
|
|
|
|
|
; A list would be written as just (1 2 3), but we have to quote
|
|
|
|
; it to stop the reader thinking it's a function.
|
|
|
|
; Also, (list 1 2 3) is the same as '(1 2 3)
|
|
|
|
|
2013-07-01 21:50:59 +00:00
|
|
|
; "Collections" are just groups of data
|
2013-06-26 17:58:50 +00:00
|
|
|
; Both lists and vectors are collections:
|
|
|
|
(coll? '(1 2 3)) ; => true
|
|
|
|
(coll? [1 2 3]) ; => true
|
|
|
|
|
2013-07-01 21:50:59 +00:00
|
|
|
; "Sequences" (seqs) are abstract descriptions of lists of data.
|
2013-06-26 17:58:50 +00:00
|
|
|
; Only lists are seqs.
|
|
|
|
(seq? '(1 2 3)) ; => true
|
|
|
|
(seq? [1 2 3]) ; => false
|
|
|
|
|
2013-07-01 21:50:59 +00:00
|
|
|
; A seq need only provide an entry when it is accessed.
|
|
|
|
; So, seqs which can be lazy -- they can define infinite series:
|
2013-06-26 17:58:50 +00:00
|
|
|
(range 4) ; => (0 1 2 3)
|
|
|
|
(range) ; => (0 1 2 3 4 ...) (an infinite series)
|
|
|
|
(take 4 (range)) ; (0 1 2 3)
|
|
|
|
|
|
|
|
; Use cons to add an item to the beginning of a list or vector
|
|
|
|
(cons 4 [1 2 3]) ; => (4 1 2 3)
|
|
|
|
(cons 4 '(1 2 3)) ; => (4 1 2 3)
|
|
|
|
|
2013-07-01 21:50:59 +00:00
|
|
|
; Conj will add an item to a collection in the most efficient way.
|
|
|
|
; For lists, they insert at the beginning. For vectors, they insert at the end.
|
2013-06-26 17:58:50 +00:00
|
|
|
(conj [1 2 3] 4) ; => [1 2 3 4]
|
|
|
|
(conj '(1 2 3) 4) ; => (4 1 2 3)
|
|
|
|
|
|
|
|
; Use concat to add lists or vectors together
|
|
|
|
(concat [1 2] '(3 4)) ; => (1 2 3 4)
|
|
|
|
|
|
|
|
; Use filter, map to interact with collections
|
|
|
|
(map inc [1 2 3]) ; => (2 3 4)
|
|
|
|
(filter even? [1 2 3]) ; => (2)
|
|
|
|
|
|
|
|
; Use reduce to reduce them
|
|
|
|
(reduce + [1 2 3 4])
|
|
|
|
; = (+ (+ (+ 1 2) 3) 4)
|
|
|
|
; => 10
|
|
|
|
|
|
|
|
; Reduce can take an initial-value argument too
|
|
|
|
(reduce conj [] '(3 2 1))
|
|
|
|
; = (conj (conj (conj [] 3) 2) 1)
|
|
|
|
; => [3 2 1]
|
|
|
|
|
|
|
|
; Functions
|
|
|
|
;;;;;;;;;;;;;;;;;;;;;
|
|
|
|
|
|
|
|
; Use fn to create new functions. A function always returns
|
|
|
|
; its last statement.
|
|
|
|
(fn [] "Hello World") ; => fn
|
|
|
|
|
|
|
|
; (You need extra parens to call it)
|
|
|
|
((fn [] "Hello World")) ; => "Hello World"
|
|
|
|
|
|
|
|
; You can create a var using def
|
|
|
|
(def x 1)
|
|
|
|
x ; => 1
|
|
|
|
|
|
|
|
; Assign a function to a var
|
|
|
|
(def hello-world (fn [] "Hello World"))
|
|
|
|
(hello-world) ; => "Hello World"
|
|
|
|
|
|
|
|
; You can shorten this process by using defn
|
|
|
|
(defn hello-world [] "Hello World")
|
|
|
|
|
|
|
|
; The [] is the list of arguments for the function.
|
|
|
|
(defn hello [name]
|
|
|
|
(str "Hello " name))
|
|
|
|
(hello "Steve") ; => "Hello Steve"
|
|
|
|
|
|
|
|
; You can also use this shorthand to create functions:
|
|
|
|
(def hello2 #(str "Hello " %1))
|
2018-09-29 00:02:14 +00:00
|
|
|
(hello2 "Julie") ; => "Hello Julie"
|
2013-06-26 17:58:50 +00:00
|
|
|
|
|
|
|
; You can have multi-variadic functions, too
|
|
|
|
(defn hello3
|
|
|
|
([] "Hello World")
|
|
|
|
([name] (str "Hello " name)))
|
|
|
|
(hello3 "Jake") ; => "Hello Jake"
|
|
|
|
(hello3) ; => "Hello World"
|
|
|
|
|
|
|
|
; Functions can pack extra arguments up in a seq for you
|
|
|
|
(defn count-args [& args]
|
|
|
|
(str "You passed " (count args) " args: " args))
|
|
|
|
(count-args 1 2 3) ; => "You passed 3 args: (1 2 3)"
|
|
|
|
|
|
|
|
; You can mix regular and packed arguments
|
|
|
|
(defn hello-count [name & args]
|
|
|
|
(str "Hello " name ", you passed " (count args) " extra args"))
|
|
|
|
(hello-count "Finn" 1 2 3)
|
|
|
|
; => "Hello Finn, you passed 3 extra args"
|
|
|
|
|
|
|
|
|
2013-07-01 21:50:59 +00:00
|
|
|
; Maps
|
2013-06-26 17:58:50 +00:00
|
|
|
;;;;;;;;;;
|
|
|
|
|
2013-07-01 21:50:59 +00:00
|
|
|
; Hash maps and array maps share an interface. Hash maps have faster lookups
|
|
|
|
; but don't retain key order.
|
2013-06-26 17:58:50 +00:00
|
|
|
(class {:a 1 :b 2 :c 3}) ; => clojure.lang.PersistentArrayMap
|
2013-07-01 21:50:59 +00:00
|
|
|
(class (hash-map :a 1 :b 2 :c 3)) ; => clojure.lang.PersistentHashMap
|
|
|
|
|
|
|
|
; Arraymaps will automatically become hashmaps through most operations
|
|
|
|
; if they get big enough, so you don't need to worry.
|
2013-06-26 17:58:50 +00:00
|
|
|
|
2013-07-01 21:50:59 +00:00
|
|
|
; Maps can use any hashable type as a key, but usually keywords are best
|
2013-06-26 17:58:50 +00:00
|
|
|
; Keywords are like strings with some efficiency bonuses
|
|
|
|
(class :a) ; => clojure.lang.Keyword
|
|
|
|
|
2013-07-01 21:50:59 +00:00
|
|
|
(def stringmap {"a" 1, "b" 2, "c" 3})
|
2013-06-26 17:58:50 +00:00
|
|
|
stringmap ; => {"a" 1, "b" 2, "c" 3}
|
|
|
|
|
2013-07-01 21:50:59 +00:00
|
|
|
(def keymap {:a 1, :b 2, :c 3})
|
|
|
|
keymap ; => {:a 1, :c 3, :b 2}
|
2013-06-26 17:58:50 +00:00
|
|
|
|
|
|
|
; By the way, commas are always treated as whitespace and do nothing.
|
|
|
|
|
|
|
|
; Retrieve a value from a map by calling it as a function
|
|
|
|
(stringmap "a") ; => 1
|
|
|
|
(keymap :a) ; => 1
|
|
|
|
|
|
|
|
; Keywords can be used to retrieve their value from a map, too!
|
|
|
|
(:b keymap) ; => 2
|
|
|
|
|
|
|
|
; Don't try this with strings.
|
|
|
|
;("a" stringmap)
|
|
|
|
; => Exception: java.lang.String cannot be cast to clojure.lang.IFn
|
|
|
|
|
2013-10-03 21:40:38 +00:00
|
|
|
; Retrieving a non-present key returns nil
|
2013-06-26 17:58:50 +00:00
|
|
|
(stringmap "d") ; => nil
|
|
|
|
|
|
|
|
; Use assoc to add new keys to hash-maps
|
2013-07-01 21:50:59 +00:00
|
|
|
(def newkeymap (assoc keymap :d 4))
|
|
|
|
newkeymap ; => {:a 1, :b 2, :c 3, :d 4}
|
2013-06-26 17:58:50 +00:00
|
|
|
|
|
|
|
; But remember, clojure types are immutable!
|
|
|
|
keymap ; => {:a 1, :b 2, :c 3}
|
|
|
|
|
|
|
|
; Use dissoc to remove keys
|
|
|
|
(dissoc keymap :a :b) ; => {:c 3}
|
|
|
|
|
|
|
|
; Sets
|
|
|
|
;;;;;;
|
|
|
|
|
|
|
|
(class #{1 2 3}) ; => clojure.lang.PersistentHashSet
|
|
|
|
(set [1 2 3 1 2 3 3 2 1 3 2 1]) ; => #{1 2 3}
|
|
|
|
|
|
|
|
; Add a member with conj
|
|
|
|
(conj #{1 2 3} 4) ; => #{1 2 3 4}
|
|
|
|
|
|
|
|
; Remove one with disj
|
|
|
|
(disj #{1 2 3} 1) ; => #{2 3}
|
|
|
|
|
|
|
|
; Test for existence by using the set as a function:
|
|
|
|
(#{1 2 3} 1) ; => 1
|
|
|
|
(#{1 2 3} 4) ; => nil
|
|
|
|
|
|
|
|
; There are more functions in the clojure.sets namespace.
|
|
|
|
|
|
|
|
; Useful forms
|
|
|
|
;;;;;;;;;;;;;;;;;
|
|
|
|
|
|
|
|
; Logic constructs in clojure are just macros, and look like
|
|
|
|
; everything else
|
|
|
|
(if false "a" "b") ; => "b"
|
|
|
|
(if false "a") ; => nil
|
|
|
|
|
|
|
|
; Use let to create temporary bindings
|
|
|
|
(let [a 1 b 2]
|
|
|
|
(> a b)) ; => false
|
|
|
|
|
|
|
|
; Group statements together with do
|
|
|
|
(do
|
|
|
|
(print "Hello")
|
|
|
|
"World") ; => "World" (prints "Hello")
|
|
|
|
|
|
|
|
; Functions have an implicit do
|
|
|
|
(defn print-and-say-hello [name]
|
|
|
|
(print "Saying hello to " name)
|
|
|
|
(str "Hello " name))
|
|
|
|
(print-and-say-hello "Jeff") ;=> "Hello Jeff" (prints "Saying hello to Jeff")
|
|
|
|
|
|
|
|
; So does let
|
|
|
|
(let [name "Urkel"]
|
|
|
|
(print "Saying hello to " name)
|
|
|
|
(str "Hello " name)) ; => "Hello Urkel" (prints "Saying hello to Urkel")
|
|
|
|
|
2015-10-17 16:31:19 +00:00
|
|
|
|
|
|
|
; Use the threading macros (-> and ->>) to express transformations of
|
|
|
|
; data more clearly.
|
|
|
|
|
|
|
|
; The "Thread-first" macro (->) inserts into each form the result of
|
|
|
|
; the previous, as the first argument (second item)
|
|
|
|
(->
|
|
|
|
{:a 1 :b 2}
|
|
|
|
(assoc :c 3) ;=> (assoc {:a 1 :b 2} :c 3)
|
|
|
|
(dissoc :b)) ;=> (dissoc (assoc {:a 1 :b 2} :c 3) :b)
|
|
|
|
|
|
|
|
; This expression could be written as:
|
|
|
|
; (dissoc (assoc {:a 1 :b 2} :c 3) :b)
|
|
|
|
; and evaluates to {:a 1 :c 3}
|
|
|
|
|
|
|
|
; The double arrow does the same thing, but inserts the result of
|
|
|
|
; each line at the *end* of the form. This is useful for collection
|
|
|
|
; operations in particular:
|
|
|
|
(->>
|
|
|
|
(range 10)
|
|
|
|
(map inc) ;=> (map inc (range 10)
|
|
|
|
(filter odd?) ;=> (filter odd? (map inc (range 10))
|
|
|
|
(into [])) ;=> (into [] (filter odd? (map inc (range 10)))
|
|
|
|
; Result: [1 3 5 7 9]
|
|
|
|
|
2016-10-02 11:55:49 +00:00
|
|
|
; When you are in a situation where you want more freedom as where to
|
|
|
|
; put the result of previous data transformations in an
|
|
|
|
; expression, you can use the as-> macro. With it, you can assign a
|
2017-04-29 08:11:05 +00:00
|
|
|
; specific name to transformations' output and use it as a
|
2016-10-02 11:55:49 +00:00
|
|
|
; placeholder in your chained expressions:
|
|
|
|
|
|
|
|
(as-> [1 2 3] input
|
|
|
|
(map inc input);=> You can use last transform's output at the last position
|
2016-10-23 12:19:13 +00:00
|
|
|
(nth input 2) ;=> and at the second position, in the same expression
|
2021-01-15 10:31:24 +00:00
|
|
|
(conj [4 5 6] input 8 9 10)) ;=> or in the middle !
|
|
|
|
; Result: [4 5 6 4 8 9 10]
|
2016-10-02 11:55:49 +00:00
|
|
|
|
|
|
|
|
2013-06-26 17:58:50 +00:00
|
|
|
; Modules
|
|
|
|
;;;;;;;;;;;;;;;
|
|
|
|
|
|
|
|
; Use "use" to get all functions from the module
|
|
|
|
(use 'clojure.set)
|
|
|
|
|
|
|
|
; Now we can use set operations
|
|
|
|
(intersection #{1 2 3} #{2 3 4}) ; => #{2 3}
|
|
|
|
(difference #{1 2 3} #{2 3 4}) ; => #{1}
|
|
|
|
|
|
|
|
; You can choose a subset of functions to import, too
|
|
|
|
(use '[clojure.set :only [intersection]])
|
|
|
|
|
|
|
|
; Use require to import a module
|
|
|
|
(require 'clojure.string)
|
|
|
|
|
|
|
|
; Use / to call functions from a module
|
2013-07-01 21:50:59 +00:00
|
|
|
; Here, the module is clojure.string and the function is blank?
|
2013-06-26 17:58:50 +00:00
|
|
|
(clojure.string/blank? "") ; => true
|
|
|
|
|
|
|
|
; You can give a module a shorter name on import
|
|
|
|
(require '[clojure.string :as str])
|
|
|
|
(str/replace "This is a test." #"[a-o]" str/upper-case) ; => "THIs Is A tEst."
|
|
|
|
; (#"" denotes a regular expression literal)
|
|
|
|
|
|
|
|
; You can use require (and use, but don't) from a namespace using :require.
|
|
|
|
; You don't need to quote your modules if you do it this way.
|
|
|
|
(ns test
|
|
|
|
(:require
|
|
|
|
[clojure.string :as str]
|
|
|
|
[clojure.set :as set]))
|
|
|
|
|
|
|
|
; Java
|
|
|
|
;;;;;;;;;;;;;;;;;
|
|
|
|
|
|
|
|
; Java has a huge and useful standard library, so
|
|
|
|
; you'll want to learn how to get at it.
|
|
|
|
|
|
|
|
; Use import to load a java module
|
|
|
|
(import java.util.Date)
|
|
|
|
|
|
|
|
; You can import from an ns too.
|
|
|
|
(ns test
|
|
|
|
(:import java.util.Date
|
|
|
|
java.util.Calendar))
|
|
|
|
|
|
|
|
; Use the class name with a "." at the end to make a new instance
|
|
|
|
(Date.) ; <a date object>
|
|
|
|
|
|
|
|
; Use . to call methods. Or, use the ".method" shortcut
|
|
|
|
(. (Date.) getTime) ; <a timestamp>
|
|
|
|
(.getTime (Date.)) ; exactly the same thing.
|
|
|
|
|
|
|
|
; Use / to call static methods
|
|
|
|
(System/currentTimeMillis) ; <a timestamp> (system is always present)
|
|
|
|
|
|
|
|
; Use doto to make dealing with (mutable) classes more tolerable
|
|
|
|
(import java.util.Calendar)
|
|
|
|
(doto (Calendar/getInstance)
|
|
|
|
(.set 2000 1 1 0 0 0)
|
|
|
|
.getTime) ; => A Date. set to 2000-01-01 00:00:00
|
2013-07-01 21:50:59 +00:00
|
|
|
|
|
|
|
; STM
|
|
|
|
;;;;;;;;;;;;;;;;;
|
|
|
|
|
|
|
|
; Software Transactional Memory is the mechanism clojure uses to handle
|
|
|
|
; persistent state. There are a few constructs in clojure that use this.
|
|
|
|
|
|
|
|
; An atom is the simplest. Pass it an initial value
|
|
|
|
(def my-atom (atom {}))
|
|
|
|
|
|
|
|
; Update an atom with swap!.
|
|
|
|
; swap! takes a function and calls it with the current value of the atom
|
|
|
|
; as the first argument, and any trailing arguments as the second
|
|
|
|
(swap! my-atom assoc :a 1) ; Sets my-atom to the result of (assoc {} :a 1)
|
|
|
|
(swap! my-atom assoc :b 2) ; Sets my-atom to the result of (assoc {:a 1} :b 2)
|
|
|
|
|
2013-10-03 21:31:56 +00:00
|
|
|
; Use '@' to dereference the atom and get the value
|
2013-07-01 21:50:59 +00:00
|
|
|
my-atom ;=> Atom<#...> (Returns the Atom object)
|
|
|
|
@my-atom ; => {:a 1 :b 2}
|
|
|
|
|
|
|
|
; Here's a simple counter using an atom
|
|
|
|
(def counter (atom 0))
|
|
|
|
(defn inc-counter []
|
|
|
|
(swap! counter inc))
|
|
|
|
|
|
|
|
(inc-counter)
|
|
|
|
(inc-counter)
|
|
|
|
(inc-counter)
|
|
|
|
(inc-counter)
|
|
|
|
(inc-counter)
|
|
|
|
|
|
|
|
@counter ; => 5
|
|
|
|
|
|
|
|
; Other STM constructs are refs and agents.
|
|
|
|
; Refs: http://clojure.org/refs
|
|
|
|
; Agents: http://clojure.org/agents
|
2013-06-26 17:58:50 +00:00
|
|
|
```
|
2013-07-01 21:50:59 +00:00
|
|
|
|
|
|
|
### Further Reading
|
|
|
|
|
2013-08-14 14:58:19 +00:00
|
|
|
This is far from exhaustive, but hopefully it's enough to get you on your feet.
|
2013-07-01 21:50:59 +00:00
|
|
|
|
|
|
|
Clojure.org has lots of articles:
|
|
|
|
[http://clojure.org/](http://clojure.org/)
|
|
|
|
|
|
|
|
Clojuredocs.org has documentation with examples for most core functions:
|
|
|
|
[http://clojuredocs.org/quickref/Clojure%20Core](http://clojuredocs.org/quickref/Clojure%20Core)
|
|
|
|
|
|
|
|
4Clojure is a great way to build your clojure/FP skills:
|
|
|
|
[http://www.4clojure.com/](http://www.4clojure.com/)
|
|
|
|
|
2013-10-28 05:42:05 +00:00
|
|
|
Clojure-doc.org (yes, really) has a number of getting started articles:
|
2013-07-01 21:50:59 +00:00
|
|
|
[http://clojure-doc.org/](http://clojure-doc.org/)
|
2019-10-24 16:29:43 +00:00
|
|
|
|
|
|
|
Clojure for the Brave and True has a great introduction to Clojure and a free online version:
|
|
|
|
[https://www.braveclojure.com/clojure-for-the-brave-and-true/](https://www.braveclojure.com/clojure-for-the-brave-and-true/)
|