LC 232. 用栈实现队列

题目描述

这是 LeetCode 上的 232. 用栈实现队列 ,难度为 简单

请你仅使用两个栈实现先入先出队列。队列应当支持一般队列支持的所有操作(push、pop、peek、empty):

实现 MyQueue 类:

  • void push(int x) 将元素 x 推到队列的末尾
  • int pop() 从队列的开头移除并返回元素
  • int peek() 返回队列开头的元素
  • boolean empty() 如果队列为空,返回 true ;否则,返回 false

说明:

  • 你只能使用标准的栈操作 —— 也就是只有 push to top, peek/pop from top, size, 和 is empty 操作是合法的。
  • 你所使用的语言也许不支持栈。你可以使用 list 或者 deque(双端队列)来模拟一个栈,只要是标准的栈操作即可。

进阶:

  • 你能否实现每个操作均摊时间复杂度为 O(1) 的队列?换句话说,执行 n 个操作的总时间复杂度为 O(n) ,即使其中一个操作可能花费较长时间。

示例:

1
2
3
4
5
6
7
8
9
10
11
12
13
输入:
["MyQueue", "push", "push", "peek", "pop", "empty"]
[[], [1], [2], [], [], []]
输出:
[null, null, null, 1, 1, false]

解释:
MyQueue myQueue = new MyQueue();
myQueue.push(1); // queue is: [1]
myQueue.push(2); // queue is: [1, 2] (leftmost is front of the queue)
myQueue.peek(); // return 1
myQueue.pop(); // return 1, queue is [2]
myQueue.empty(); // return false

提示:

  • 1 <= x <= 9
  • 最多调用 100 次 push、pop、peek 和 empty
  • 假设所有操作都是有效的 (例如,一个空的队列不会调用 pop 或者 peek 操作)

基本思路

无论「用栈实现队列」还是「用队列实现栈」,思路都是类似的。

都可以通过使用两个栈/队列来解决。

我们创建两个栈,分别为 outin,用作处理「输出」和「输入」操作。

其实就是两个栈来回「倒腾」。

而对于「何时倒腾」决定了是 O(n) 解法 还是 均摊 O(1) 解法


O(n) 解法

我们创建两个栈,分别为 outin

  • in 用作处理输入操作 push(),使用 in 时需确保 out 为空
  • out 用作处理输出操作 pop()peek(),使用 out 时需确保 in 为空
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
class MyQueue {
Deque<Integer> out, in;
public MyQueue() {
in = new ArrayDeque<>();
out = new ArrayDeque<>();
}

public void push(int x) {
while (!out.isEmpty()) in.addLast(out.pollLast());
in.addLast(x);
}

public int pop() {
while (!in.isEmpty()) out.addLast(in.pollLast());
return out.pollLast();
}

public int peek() {
while (!in.isEmpty()) out.addLast(in.pollLast());
return out.peekLast();
}

public boolean empty() {
return out.isEmpty() && in.isEmpty();
}
}
  • 时间复杂度:$O(n)$
  • 空间复杂度:$O(n)$

均摊 O(1) 解法

事实上,我们不需要在每次的「入栈」和「出栈」操作中都进行「倒腾」。

我们只需要保证,输入的元素总是跟在前面的输入元素的后面,而输出元素总是最早输入的那个元素即可。

可以通过调整「倒腾」的时机来确保满足上述要求,但又不需要发生在每一次操作中:

  • 只有在「输出栈」为空的时候,才发生一次性的「倒腾」
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
class MyQueue {
Deque<Integer> out, in;
public MyQueue() {
in = new ArrayDeque<>();
out = new ArrayDeque<>();
}

public void push(int x) {
in.addLast(x);
}

public int pop() {
if (out.isEmpty()) {
while (!in.isEmpty()) out.addLast(in.pollLast());
}
return out.pollLast();
}

public int peek() {
if (out.isEmpty()) {
while (!in.isEmpty()) out.addLast(in.pollLast());
}
return out.peekLast();
}

public boolean empty() {
return out.isEmpty() && in.isEmpty();
}
}
  • 时间复杂度:pop()peek() 操作都是均摊 $O(1)$
  • 空间复杂度:$O(n)$

关于「均摊复杂度」的说明

我们先用另外一个例子来理解「均摊复杂度」,大家都知道「哈希表」底层是通过数组实现的。

正常情况下,计算元素在哈希桶的位置,然后放入哈希桶,复杂度为 $O(1)$,假定是通过简单的“拉链法”搭配「头插法」方式来解决哈希冲突。

但当某次元素插入后,「哈希表」达到扩容阈值,则需要对底层所使用的数组进行扩容,这个复杂度是 $O(n)$

显然「扩容」操作不会发生在每一次的元素插入中,因此扩容的 $O(n)$ 都会伴随着 n 次的 $O(1)$,也就是 $O(n)$ 的复杂度会被均摊到每一次插入当中,因此哈希表插入仍然是 $O(1)$ 的。

同理,我们的「倒腾」不是发生在每一次的「输出操作」中,而是集中发生在一次「输出栈为空」的时候,因此 poppeek 都是均摊复杂度为 $O(1)$ 的操作。

由于本题的调用次数只有 100 次,所以铁定是一个人均 100% 的算法(0 ms)🐶 🐶

我们需要对操作进行复杂度分析进行判断,而不是看时间来判断自己是不是均摊 O(1) 哦 ~


最后

这是我们「刷穿 LeetCode」系列文章的第 No.232 篇,系列开始于 2021/01/01,截止于起始日 LeetCode 上共有 1916 道题目,部分是有锁题,我们将先把所有不带锁的题目刷完。

在这个系列文章里面,除了讲解解题思路以外,还会尽可能给出最为简洁的代码。如果涉及通解还会相应的代码模板。

为了方便各位同学能够电脑上进行调试和提交代码,我建立了相关的仓库:https://github.com/SharingSource/LogicStack-LeetCode

在仓库地址里,你可以看到系列文章的题解链接、系列文章的相应代码、LeetCode 原题链接和其他优选题解。


本博客所有文章除特别声明外,均采用 CC BY-SA 4.0 协议 ,转载请注明出处!