Immutable data cannot be changed once created, leading to much simpler
application development and enabling techniques from functional programming such
as lazy evaluation. Immutable JS provides a lazy Sequence
, allowing efficient
chaining of sequence methods like map
and filter
without creating
intermediate representations.
immutable
provides Sequence
, Range
, Repeat
, Map
, OrderedMap
, Set
and a sparse Vector
by using lazy sequences and hash maps tries.
They achieve efficiency by using structural sharing and minimizing the need to
copy or cache data.
Install immutable
using npm
npm install immutable
Then require it into any module.
var Immutable = require('immutable');
var map = Immutable.Map({a:1, b:2, c:3});
To use immutable
from a browser, try Browserify.
Use these Immutable collections and sequences as you would use native collections in your TypeScript programs while still taking advantage of type generics, error detection, and auto-complete in your IDE.
(Because of TypeScript 1.0's issue with NodeJS module resolution, you must require the full file path)
import Immutable = require('./node_modules/immutable/dist/Immutable');
var map: Immutable.Map<string, number>;
map = Immutable.Map({a:1, b:2, c:3});
map = map.set('b', 20);
map.get('b'); // 20
Much of what makes application development difficult is tracking mutation and maintaining state. Developing with immutable data encourages you to think differently about how data flows through your application.
Subscribing to data events throughout your application, by using
Object.observe
, or any other mechanism, creates a huge overhead of
book-keeping which can hurt performance, sometimes dramatically, and creates
opportunities for areas of your application to get out of sync with each other
due to simple programmer error. Since immutable data never changes, subscribing
to changes throughout the model is a dead-end and new data can only ever be
passed from above.
This model of data flow aligns well with the architecture of React and especially well with an application designed using the ideas of Flux.
When data is passed from above rather than being subscribed to, and you're only
interested in doing work when something has changed, you can use equality.
immutable
always returns itself when a mutation results in an identical
collection, allowing for using ===
equality to determine if something
has changed.
var map1 = Immutable.Map({a:1, b:2, c:3});
var map2 = map1.set('b', 2);
assert(map1 === map2);
If an object is immutable, it can be "cloned" simply by making another reference to it instead of copying the entire object. Because a reference is much smaller than the object itself, this results in memory savings and a potential boost in execution speed for programs which rely on copies (such as an undo-stack).
var map1 = Immutable.Map({a:1, b:2, c:3});
var clone = map1;
While immutable
is inspired by Clojure, Haskell and other functional
programming environments, it's designed to bring these powerful concepts to
JavaScript, and therefore has an Object-Oriented API that closely mirrors that
of Array,
Map, and
Set.
The difference for the immutable collections is that methods which would mutate
the collection, like push
, set
, unshift
or splice
instead return a new
immutable collection. Methods which return new arrays like slice
or concat
instead return new immutable collections.
var vect1 = Immutable.Vector(1, 2);
var vect2 = vect1.push(3, 4, 5);
var vect3 = vect2.unshift(0);
var vect4 = vect1.concat(vect2, vect3);
assert(vect1.length === 2);
assert(vect2.length === 5);
assert(vect3.length === 6);
assert(vect4.length === 13);
assert(vect4.get(0) === 1);
Almost all of the methods on Array
will be found in similar form on
Immutable.Vector
, those of Map
found on Immutable.Map
, and those of Set
found on Immutable.Set
, including sequence operations like forEach
and map
.
var alpha = Immutable.Map({a:1, b:2, c:3, d:4});
alpha.map((v, k) => k.toUpperCase()).join();
// 'A,B,C,D'
Designed to inter-operate with your existing JavaScript, immutable
accepts plain JavaScript Arrays and Objects anywhere a method expects a
Sequence
with no performance penalty.
var map1 = Immutable.Map({a:1, b:2, c:3, d:4});
var map2 = Immutable.Map({c:10, a:20, t:30});
var obj = {d:100, o:200, g:300};
var map3 = map1.merge(map2, obj);
// Map { a: 20, b: 2, c: 10, d: 100, t: 30, o: 200, g: 300 }
This is possible because immutable
can treat any JavaScript Array or Object
as a Sequence. You can take advantage of this in order to get sophisticated
sequence methods on JavaScript Objects, which otherwise have a very sparse
native API. Because Sequences evaluate lazily and do not cache intermediate
results, these operations are extremely efficient.
var myObject = {a:1,b:2,c:3};
Sequence(myObject).map(x => x * x).toObject();
// { a: 1, b: 4, c: 9 }
All immutable
Sequences can be converted to plain JavaScript Arrays and
Objects shallowly with toArray()
and toObject()
or deeply with toJS()
.
All sequences also implement toJSON()
allowing them to be passed to
JSON.stringify
directly.
var deep = Immutable.Map({a:1, b:2, c:Immutable.Vector(3,4,5)});
deep.toObject() // { a: 1, b: 2, c: Vector [ 3, 4, 5 ] }
deep.toArray() // [ 1, 2, Vector [ 3, 4, 5 ] ]
deep.toJS() // { a: 1, b: 2, c: [ 3, 4, 5 ] }
JSON.stringify(deep) // '{"a":1,"b":2,"c":[3,4,5]}'
The collections in immutable
are intended to be nested, allowing for deep
trees of data, similar to JSON.
var nested = Immutable.fromJS({a:{b:{c:[3,4,5]}}});
// Map { a: Map { b: Map { c: Vector [ 3, 4, 5 ] } } }
A few power-tools allow for reading and operating on nested data. The
most useful are mergeDeep
, getIn
and updateIn
, found on Vector
, Map
and OrderedMap
.
var nested2 = nested.mergeDeep({a:{b:{d:6}}});
// Map { a: Map { b: Map { c: Vector [ 3, 4, 5 ], d: 6 } } }
nested2.getIn(['a', 'b', 'd']); // 6
var nested3 = nested2.updateIn(['a', 'b', 'd'], value => value + 1);
// Map { a: Map { b: Map { c: Vector [ 3, 4, 5 ], d: 7 } } }
The Sequence
is a set of (key, value) entries which can be iterated, and
is the base class for all collections in immutable
, allowing them to make
use of all the Sequence methods (such as map
and filter
).
Sequences are immutable — Once a sequence is created, it cannot be changed, appended to, rearranged or otherwise modified. Instead, any mutative method called on a sequence will return a new immutable sequence.
Sequences are lazy — Sequences do as little work as necessary to respond to any method call.
For example, the following does no work, because the resulting sequence is never used:
var oddSquares = Immutable.Sequence(1,2,3,4,5,6,7,8)
.filter(x => x % 2).map(x => x * x);
Once the sequence is used, it performs only the work necessary. In this example, no intermediate arrays are ever created, filter is only called twice, and map is only called once:
console.log(oddSquares.last()); // 49
Lazy Sequences allow for the efficient chaining of sequence operations, allowing for the expression of logic that can otherwise be very tedious:
Immutable.Sequence({a:1, b:1, c:1})
.flip().map(key => key.toUpperCase()).flip().toObject();
// Map { A: 1, B: 1, C: 1 }
As well as expressing logic that would otherwise seem memory-limited:
Immutable.Range(1, Infinity)
.skip(1000)
.map(n => -n)
.filter(n => n % 2 === 0)
.take(2)
.reduce((r, n) => r * n, 1);
// 1006008
Note: A sequence is always iterated in the same order, however that order may
not always be well defined, as is the case for the Map
.
immutable
provides equality which treats immutable data structures as
pure data, performing a deep equality check if necessary.
var map1 = Immutable.Map({a:1, b:1, c:1});
var map2 = Immutable.Map({a:1, b:1, c:1});
assert(map1 !== map2);
assert(Immutable.is(map1, map2) === true);
Immutable.is
uses the same measure of equality as Object.is
including if both are immutable sequences and all keys and values are equal
using the same measure of equality.
If a tree falls in the woods, does it make a sound?
If a pure function mutates some local data in order to produce an immutable return value, is that ok?
— Rich Hickey, Clojure
There is a performance penalty paid every time you create a new immutable object
via applying a mutation. If you need to apply a series of mutations
immutable
gives you the ability to create a temporary mutable copy of a
collection and applying a batch of mutations in a highly performant manner by
using withMutations
. In fact, this is exactly how immutable
applies
complex mutations itself.
As an example, this results in the creation of 2, not 4, new immutable Vectors.
var vect1 = Immutable.Vector(1,2,3);
var vect2 = vect1.withMutations(function (vect) {
vect.push(4).push(5).push(6);
});
assert(vect1.length === 3);
assert(vect2.length === 6);
Note: immutable
also provides asMutable
and asImmutable
, but only
encourages their use when withMutations
will not suffice.
All documentation is contained within the type definition file, Immutable.d.ts.
Use Github issues for requests.
We actively welcome pull requests, learn how to contribute.
Hugh Jackson, for providing the npm package name. If you're looking for his unsupported package, see v1.4.1.
immutable
is BSD-licensed. We also provide an additional patent grant.