列表

详情


295. 数据流的中位数

中位数是有序列表中间的数。如果列表长度是偶数,中位数则是中间两个数的平均值。

例如,

[2,3,4] 的中位数是 3

[2,3] 的中位数是 (2 + 3) / 2 = 2.5

设计一个支持以下两种操作的数据结构:

示例:

addNum(1)
addNum(2)
findMedian() -> 1.5
addNum(3) 
findMedian() -> 2

进阶:

  1. 如果数据流中所有整数都在 0 到 100 范围内,你将如何优化你的算法?
  2. 如果数据流中 99% 的整数都在 0 到 100 范围内,你将如何优化你的算法?

相似题目

滑动窗口中位数

原站题解

去查看

上次编辑到这里,代码来自缓存 点击恢复默认模板
class MedianFinder { public: MedianFinder() { } void addNum(int num) { } double findMedian() { } }; /** * Your MedianFinder object will be instantiated and called as such: * MedianFinder* obj = new MedianFinder(); * obj->addNum(num); * double param_2 = obj->findMedian(); */

php 解法, 执行用时: 416 ms, 内存消耗: 63.1 MB, 提交时间: 2023-09-11 11:04:16

class MedianFinder {
    private $count;
    private $minHeap;
    private $maxHeap;
    
    /**
     * initialize your data structure here.
     */
    function __construct() {
        $this->count = 0;
        $this->minHeap = new SplMinHeap;
        $this->maxHeap = new SplMaxHeap;
    }

    /**
     * @param Integer $num
     * @return NULL
     */
    function addNum($num) {
        $this->count += 1;
        
        $this->maxHeap->insert($num);
        $this->minHeap->insert($this->maxHeap->extract());
        if (($this->count & 1) != 0) {
            $this->maxHeap->insert($this->minHeap->extract());
        }    
    }
  
    /**
     * @return Float
     */
    function findMedian() {
        if (($this->count & 1) == 0) {
            return (double) ($this->maxHeap->top() + $this->minHeap->top()) / 2;
        } else {
            return (double) $this->maxHeap->top();
        }
    }
}

/**
 * Your MedianFinder object will be instantiated and called as such:
 * $obj = MedianFinder();
 * $obj->addNum($num);
 * $ret_2 = $obj->findMedian();
 */

golang 解法, 执行用时: 320 ms, 内存消耗: 20.7 MB, 提交时间: 2023-09-11 11:03:10

type MedianFinder struct {
    nums        *redblacktree.Tree
    total       int
    left, right iterator
}

func Constructor() MedianFinder {
    return MedianFinder{nums: redblacktree.NewWithIntComparator()}
}

func (mf *MedianFinder) AddNum(num int) {
    if count, has := mf.nums.Get(num); has {
        mf.nums.Put(num, count.(int)+1)
    } else {
        mf.nums.Put(num, 1)
    }
    if mf.total == 0 {
        it := mf.nums.Iterator()
        it.Next()
        mf.left = iterator{it, 1}
        mf.right = mf.left
    } else if mf.total%2 == 1 {
        if num < mf.left.Key().(int) {
            mf.left.prev()
        } else {
            mf.right.next()
        }
    } else {
        if mf.left.Key().(int) < num && num < mf.right.Key().(int) {
            mf.left.next()
            mf.right.prev()
        } else if num >= mf.right.Key().(int) {
            mf.left.next()
        } else {
            mf.right.prev()
            mf.left = mf.right
        }
    }
    mf.total++
}

func (mf *MedianFinder) FindMedian() float64 {
    return float64(mf.left.Key().(int)+mf.right.Key().(int)) / 2
}

type iterator struct {
    redblacktree.Iterator
    count int
}

func (it *iterator) prev() {
    if it.count > 1 {
        it.count--
    } else {
        it.Prev()
        it.count = it.Value().(int)
    }
}

func (it *iterator) next() {
    if it.count < it.Value().(int) {
        it.count++
    } else {
        it.Next()
        it.count = 1
    }
}

/**
 * Your MedianFinder object will be instantiated and called as such:
 * obj := Constructor();
 * obj.AddNum(num);
 * param_2 := obj.FindMedian();
 */

python3 解法, 执行用时: 964 ms, 内存消耗: 38 MB, 提交时间: 2023-09-11 11:02:46

from sortedcontainers import SortedList

class MedianFinder:

    def __init__(self):
        self.nums = SortedList()
        self.left = self.right = None
        self.left_value = self.right_value = None

    def addNum(self, num: int) -> None:
        nums_ = self.nums

        n = len(nums_)
        nums_.add(num)

        if n == 0:
            self.left = self.right = 0
        else:
            # 模拟双指针,当 num 小于 self.left 或 self.right 指向的元素时,num 的加入会导致对应指针向右移动一个位置
            if num < self.left_value:
                self.left += 1
            if num < self.right_value:
                self.right += 1

            if n & 1:
                if num < self.left_value:
                    self.left -= 1
                else:
                    self.right += 1
            else:
                if self.left_value < num < self.right_value:
                    self.left += 1
                    self.right -= 1
                elif num >= self.right_value:
                    self.left += 1
                else:
                    self.right -= 1
                    self.left = self.right
        
        self.left_value = nums_[self.left]
        self.right_value = nums_[self.right]

    def findMedian(self) -> float:
        return (self.left_value + self.right_value) / 2


# Your MedianFinder object will be instantiated and called as such:
# obj = MedianFinder()
# obj.addNum(num)
# param_2 = obj.findMedian()

java 解法, 执行用时: 152 ms, 内存消耗: 73 MB, 提交时间: 2023-09-11 11:02:27

class MedianFinder {
    TreeMap<Integer, Integer> nums;
    int n;
    int[] left;
    int[] right;

    public MedianFinder() {
        nums = new TreeMap<Integer, Integer>();
        n = 0;
        left = new int[2];
        right = new int[2];
    }
    
    public void addNum(int num) {
        nums.put(num, nums.getOrDefault(num, 0) + 1);
        if (n == 0) {
            left[0] = right[0] = num;
            left[1] = right[1] = 1;
        } else if ((n & 1) != 0) {
            if (num < left[0]) {
                decrease(left);
            } else {
                increase(right);
            }
        } else {
            if (num > left[0] && num < right[0]) {
                increase(left);
                decrease(right);
            } else if (num >= right[0]) {
                increase(left);
            } else {
                decrease(right);
                System.arraycopy(right, 0, left, 0, 2);
            }
        }
        n++;
    }

    public double findMedian() {
        return (left[0] + right[0]) / 2.0;
    }

    private void increase(int[] iterator) {
        iterator[1]++;
        if (iterator[1] > nums.get(iterator[0])) {
            iterator[0] = nums.ceilingKey(iterator[0] + 1);
            iterator[1] = 1;
        }
    }

    private void decrease(int[] iterator) {
        iterator[1]--;
        if (iterator[1] == 0) {
            iterator[0] = nums.floorKey(iterator[0] - 1);
            iterator[1] = nums.get(iterator[0]);
        }
    }
}


/**
 * Your MedianFinder object will be instantiated and called as such:
 * MedianFinder obj = new MedianFinder();
 * obj.addNum(num);
 * double param_2 = obj.findMedian();
 */

cpp 解法, 执行用时: 356 ms, 内存消耗: 124.8 MB, 提交时间: 2023-09-11 11:02:08

// 有序集合 + 双指针
class MedianFinder {
    multiset<int> nums;
    multiset<int>::iterator left, right;

public:
    MedianFinder() : left(nums.end()), right(nums.end()) {}

    void addNum(int num) {
        const size_t n = nums.size();

        nums.insert(num);
        if (!n) {
            left = right = nums.begin();
        } else if (n & 1) {
            if (num < *left) {
                left--;
            } else {
                right++;
            }
        } else {
            if (num > *left && num < *right) {
                left++;
                right--;
            } else if (num >= *right) {
                left++;
            } else {
                right--;
                left = right;
            }
        }
    }

    double findMedian() {
        return (*left + *right) / 2.0;
    }
};

/**
 * Your MedianFinder object will be instantiated and called as such:
 * MedianFinder* obj = new MedianFinder();
 * obj->addNum(num);
 * double param_2 = obj->findMedian();
 */

cpp 解法, 执行用时: 284 ms, 内存消耗: 114 MB, 提交时间: 2023-09-11 11:01:31

class MedianFinder {
public:
    priority_queue<int, vector<int>, less<int>> queMin;
    priority_queue<int, vector<int>, greater<int>> queMax;

    MedianFinder() {}

    void addNum(int num) {
        if (queMin.empty() || num <= queMin.top()) {
            queMin.push(num);
            if (queMax.size() + 1 < queMin.size()) {
                queMax.push(queMin.top());
                queMin.pop();
            }
        } else {
            queMax.push(num);
            if (queMax.size() > queMin.size()) {
                queMin.push(queMax.top());
                queMax.pop();
            }
        }
    }

    double findMedian() {
        if (queMin.size() > queMax.size()) {
            return queMin.top();
        }
        return (queMin.top() + queMax.top()) / 2.0;
    }
};

/**
 * Your MedianFinder object will be instantiated and called as such:
 * MedianFinder* obj = new MedianFinder();
 * obj->addNum(num);
 * double param_2 = obj->findMedian();
 */

java 解法, 执行用时: 97 ms, 内存消耗: 69.2 MB, 提交时间: 2023-09-11 11:01:14

class MedianFinder {
    PriorityQueue<Integer> queMin;
    PriorityQueue<Integer> queMax;

    public MedianFinder() {
        queMin = new PriorityQueue<Integer>((a, b) -> (b - a));
        queMax = new PriorityQueue<Integer>((a, b) -> (a - b));
    }
    
    public void addNum(int num) {
        if (queMin.isEmpty() || num <= queMin.peek()) {
            queMin.offer(num);
            if (queMax.size() + 1 < queMin.size()) {
                queMax.offer(queMin.poll());
            }
        } else {
            queMax.offer(num);
            if (queMax.size() > queMin.size()) {
                queMin.offer(queMax.poll());
            }
        }
    }
    
    public double findMedian() {
        if (queMin.size() > queMax.size()) {
            return queMin.peek();
        }
        return (queMin.peek() + queMax.peek()) / 2.0;
    }
}


/**
 * Your MedianFinder object will be instantiated and called as such:
 * MedianFinder obj = new MedianFinder();
 * obj.addNum(num);
 * double param_2 = obj.findMedian();
 */

golang 解法, 执行用时: 276 ms, 内存消耗: 22.4 MB, 提交时间: 2023-09-11 11:00:56

type MedianFinder struct {
    queMin, queMax hp
}

func Constructor() MedianFinder {
    return MedianFinder{}
}

func (mf *MedianFinder) AddNum(num int) {
    minQ, maxQ := &mf.queMin, &mf.queMax
    if minQ.Len() == 0 || num <= -minQ.IntSlice[0] {
        heap.Push(minQ, -num)
        if maxQ.Len()+1 < minQ.Len() {
            heap.Push(maxQ, -heap.Pop(minQ).(int))
        }
    } else {
        heap.Push(maxQ, num)
        if maxQ.Len() > minQ.Len() {
            heap.Push(minQ, -heap.Pop(maxQ).(int))
        }
    }
}

func (mf *MedianFinder) FindMedian() float64 {
    minQ, maxQ := mf.queMin, mf.queMax
    if minQ.Len() > maxQ.Len() {
        return float64(-minQ.IntSlice[0])
    }
    return float64(maxQ.IntSlice[0]-minQ.IntSlice[0]) / 2
}

type hp struct{ sort.IntSlice }
func (h *hp) Push(v interface{}) { h.IntSlice = append(h.IntSlice, v.(int)) }
func (h *hp) Pop() interface{}   { a := h.IntSlice; v := a[len(a)-1]; h.IntSlice = a[:len(a)-1]; return v }


/**
 * Your MedianFinder object will be instantiated and called as such:
 * obj := Constructor();
 * obj.AddNum(num);
 * param_2 := obj.FindMedian();
 */

python3 解法, 执行用时: 364 ms, 内存消耗: 36.8 MB, 提交时间: 2023-09-11 11:00:34

# 优先队列
import heapq

class MedianFinder:
    def __init__(self):
        self.queMin = list()
        self.queMax = list()

    def addNum(self, num: int) -> None:
        queMin_ = self.queMin
        queMax_ = self.queMax

        if not queMin_ or num <= -queMin_[0]:
            heapq.heappush(queMin_, -num)
            if len(queMax_) + 1 < len(queMin_):
                heapq.heappush(queMax_, -heapq.heappop(queMin_))
        else:
            heapq.heappush(queMax_, num)
            if len(queMax_) > len(queMin_):
                heapq.heappush(queMin_, -heapq.heappop(queMax_))
        
    def findMedian(self) -> float:
        queMin_ = self.queMin
        queMax_ = self.queMax

        if len(queMin_) > len(queMax_):
            return -queMin_[0]
        return (-queMin_[0] + queMax_[0]) / 2


# Your MedianFinder object will be instantiated and called as such:
# obj = MedianFinder()
# obj.addNum(num)
# param_2 = obj.findMedian()

上一题