使用两个堆栈构造一个队列。

O(1)用于入队和出队操作。

class Queue {
  constructor() {
    this.s1 = []; // in
    this.s2 = []; // out
  }

  enqueue(val) {
    this.s1.push(val);
  }

  dequeue() {
    if (this.s2.length === 0) {
      this._move();
    }

    return this.s2.pop(); // return undefined if empty
  }

  _move() {
    while (this.s1.length) {
      this.s2.push(this.s1.pop());
    }
  }
}

As many have said: native array using push and pop is fine for a stack, but using shift for taking elements from a queue means that the remaining elements need to move, which is potentially slow. The idea of using two stacks to make a queue in kevinyu's answer is a nice idea to fix it, and of course that can be done with native-array-stacks as well. (Edit: there was actually already an answer by Yuki-Dreamer that does this, albeit less compactly. I didn't notice it until now because it was unfairly downvoted.)

下面是一个使用ES5/ES6特性的紧凑实现，它使队列对象的行为尽可能接近本机的push/shift变体，除了每次操作花费O(1)平摊时间:

const queue = () => {
    const a = [], b = [];
    return {
        push: (...elts) => a.push(...elts),
        shift: () => {
            if (b.length === 0) {
                while (a.length > 0) { b.push(a.pop()) }
            }
            return b.pop();
        },
        get length() { return a.length + b.length }
    }
}

现在你可以做:

const q = queue();
q.push(8);
q.push(9);
q.push(10);
console.log(q.length);          // outputs 3
console.log(q.shift());         // outputs 8
q.push(11);
console.log(q.shift());         // outputs 9
console.log(q.shift());         // outputs 10
console.log(q.shift());         // outputs 11
console.log(q.shift());         // outputs undefined

队列实现对长度使用getter语法，使其看起来像一个属性，并对push使用rest参数语法，以允许一次推送多个内容。如果你不想这样做，你可以用push: elt => a.push(elt)，替换第4行。(但是请注意，你不能用push: a.push替换它，就像我自己第一次尝试的那样，结果非常奇怪:这是因为它导致本机push方法被调用，并设置为队列对象。)

我在实现BFS时遇到了这个线程。在疑惑为何表现如此糟糕之后，我做了一些调查。array.shift()通常在O(n)中运行，这将我的BFS运行时从O(V+E)增加到O(V^2+E)。

我没有从头开始实现一个队列，而是使用了npm包双端队列，它与之前使用的数组方法兼容，工作起来很有魅力。 deque可以用作堆栈或队列。

    //import package
    import Deque from 'double-ended-queue';

    //create queue
    let queue = new Deque();
    //append
    queue.push(item);
    //dequeue (get first item inserted)
    let firstItem = queue.shift();

    //pop (get last item inserted)
    let lastItem = queue.pop();

如果你理解栈的push()和pop()函数，那么queue只是在相反的意义上进行这些操作之一。push()的对边是unshift()， pop()的对边是shift()。 然后:

//classic stack
var stack = [];
stack.push("first"); // push inserts at the end
stack.push("second");
stack.push("last");
stack.pop(); //pop takes the "last" element

//One way to implement queue is to insert elements in the oposite sense than a stack
var queue = [];
queue.unshift("first"); //unshift inserts at the beginning
queue.unshift("second");
queue.unshift("last");
queue.pop(); //"first"

//other way to do queues is to take the elements in the oposite sense than stack
var queue = [];
queue.push("first"); //push, as in the stack inserts at the end
queue.push("second");
queue.push("last");
queue.shift(); //but shift takes the "first" element

Create a pair of classes that provide the various methods that each of these data structures has (push, pop, peek, etc). Now implement the methods. If you're familiar with the concepts behind stack/queue, this should be pretty straightforward. You can implement the stack with an array, and a queue with a linked list, although there are certainly other ways to go about it. Javascript will make this easy, because it is weakly typed, so you don't even have to worry about generic types, which you'd have to do if you were implementing it in Java or C#.

Javascript数组shift()很慢，特别是当包含很多元素时。我知道两种方法来实现平摊O(1)复杂度的队列。

首先是使用循环缓冲区和表加倍。我以前实现过这个。你可以在这里看到我的源代码 https://github.com/kevyuu/rapid-queue

第二种方法是使用两个堆栈。这是带有两个堆栈的队列的代码

function createDoubleStackQueue() {
var that = {};
var pushContainer = [];
var popContainer = [];

function moveElementToPopContainer() {
    while (pushContainer.length !==0 ) {
        var element = pushContainer.pop();
        popContainer.push(element);
    }
}

that.push = function(element) {
    pushContainer.push(element);
};

that.shift = function() {
    if (popContainer.length === 0) {
        moveElementToPopContainer();
    }
    if (popContainer.length === 0) {
        return null;
    } else {
        return popContainer.pop();
    }
};

that.front = function() {
    if (popContainer.length === 0) {
        moveElementToPopContainer();
    }
    if (popContainer.length === 0) {
        return null;
    }
    return popContainer[popContainer.length - 1];
};

that.length = function() {
    return pushContainer.length + popContainer.length;
};

that.isEmpty = function() {
    return (pushContainer.length + popContainer.length) === 0;
};

return that;}

这是使用jsPerf进行的性能比较

CircularQueue.shift() vs Array.shift()

http://jsperf.com/rapidqueue-shift-vs-array-shift

正如你所看到的，大数据集的速度要快得多