Given an integer array nums
, return the number of range sums that lie in [lower, upper]
inclusive. 数组
Range sum S(i, j)
is defined as the sum of the elements in nums
between indices i
and j
(i
≤ j
), inclusive.post
Note: spa
A naive algorithm of O(n2) is trivial. You MUST do better than that.指针
Example: code
Given nums = [-2, 5, -1]
, lower = -2
, upper = 2
, Return 3
. The three ranges are : [0, 0]
, [2, 2]
, [0, 2]
and their respective sums are: -2, -1, 2
.对象
这道题最直观的一个想法就是枚举出全部的子数组,而后检查他们是否在要求的取值范围内,这种方法的时间复杂度是O(n^2)的,显然会超时。blog
看到这种题目最容易想到的是什么呢?Two Pointers!对,可是在这道题上仅仅使用Two Pointers确定是不够的,在Two Pointers的思想基础上,融合归并排序,就能找到一个比较好的解决方案。排序
这里咱们的排序对象是前缀求和数组,在归并排序的合并阶段,咱们有左数组和右数组,且左和右数组都是排好序的,因此咱们能够用i遍历左数组,j,k两个指针分别取在右数组搜索,使得:three
那么此时,咱们就找到了j-k个符合要求的子数组。element
因为左右数组都是排好序的,因此当i递增以后,j和k的位置不用从头开始扫描。
最后还有一点须要注意的就是,为了防止溢出,咱们的vector容纳的是long long型元素。
class Solution { public: int countRangeSum(vector<int>& nums, int lower, int upper) { int n = nums.size(); if (n <= 0) { return 0; } vector<long long> sums(n + 1, 0); for (int i = 0; i < n; ++i) { sums[i+1] = sums[i] + nums[i]; } return merge(sums, 0, n, lower, upper); } int merge(vector<long long>& sums, int start, int end, int lower, int upper) { if (start >= end) { return 0; } int mid = start + (end - start) / 2; int count = merge(sums, start, mid, lower, upper) + merge(sums, mid + 1, end, lower, upper); vector<long long> tmp(end - start + 1, 0); int j = mid + 1, k = mid + 1, t = mid + 1, i = start, r = 0; for (; i <= mid; ++i, ++r) { while (j <= end && sums[j] - sums[i] <= upper) ++j; while (k <= end && sums[k] - sums[i] < lower) ++k; count += j - k; while (t <= end && sums[t] <= sums[i]) tmp[r++] = sums[t++]; tmp[r] = sums[i]; } for (int i = 0; i < r; ++i) { sums[start + i] = tmp[i]; } return count; } };