1749. Maximum Absolute Sum of Any Subarray

Description

You are given an integer array nums. The absolute sum of a subarray [numsl, numsl+1, ..., numsr-1, numsr] is abs(numsl + numsl+1 + ... + numsr-1 + numsr).

Return the maximum absolute sum of any (possibly empty) subarray of nums.

Note that abs(x) is defined as follows:

  • If x is a negative integer, then abs(x) = -x.
  • If x is a non-negative integer, then abs(x) = x.

 

Example 1:

Input: nums = [1,-3,2,3,-4]
Output: 5
Explanation: The subarray [2,3] has absolute sum = abs(2+3) = abs(5) = 5.

Example 2:

Input: nums = [2,-5,1,-4,3,-2]
Output: 8
Explanation: The subarray [-5,1,-4] has absolute sum = abs(-5+1-4) = abs(-8) = 8.

 

Constraints:

  • 1 <= nums.length <= 105
  • -104 <= nums[i] <= 104

Solutions

Solution 1: Dynamic Programming

We define $f[i]$ to represent the maximum value of the subarray ending with $nums[i]$, and define $g[i]$ to represent the minimum value of the subarray ending with $nums[i]$. Then the state transition equation of $f[i]$ and $g[i]$ is as follows:

$$ \begin{aligned} f[i] &= \max(f[i - 1], 0) + nums[i] \ g[i] &= \min(g[i - 1], 0) + nums[i] \end{aligned} $$

The final answer is the maximum value of $max(f[i], |g[i]|)$.

Since $f[i]$ and $g[i]$ are only related to $f[i - 1]$ and $g[i - 1]$, we can use two variables to replace the array, reducing the space complexity to $O(1)$.

Time complexity $O(n)$, space complexity $O(1)$, where $n$ is the length of the array $nums$.

Python Code
1
2
3
4
5
6
7
8
9

class Solution:
    def maxAbsoluteSum(self, nums: List[int]) -> int:
        f = g = 0
        ans = 0
        for x in nums:
            f = max(f, 0) + x
            g = min(g, 0) + x
            ans = max(ans, f, abs(g))
        return ans

Java Code
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12

class Solution {
    public int maxAbsoluteSum(int[] nums) {
        int f = 0, g = 0;
        int ans = 0;
        for (int x : nums) {
            f = Math.max(f, 0) + x;
            g = Math.min(g, 0) + x;
            ans = Math.max(ans, Math.max(f, Math.abs(g)));
        }
        return ans;
    }
}

C++ Code
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13

class Solution {
public:
    int maxAbsoluteSum(vector<int>& nums) {
        int f = 0, g = 0;
        int ans = 0;
        for (int& x : nums) {
            f = max(f, 0) + x;
            g = min(g, 0) + x;
            ans = max({ans, f, abs(g)});
        }
        return ans;
    }
};

Go Code
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16

func maxAbsoluteSum(nums []int) (ans int) {
	var f, g int
	for _, x := range nums {
		f = max(f, 0) + x
		g = min(g, 0) + x
		ans = max(ans, max(f, abs(g)))
	}
	return
}

func abs(x int) int {
	if x < 0 {
		return -x
	}
	return x
}

TypeScript Code
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11

function maxAbsoluteSum(nums: number[]): number {
    let f = 0;
    let g = 0;
    let ans = 0;
    for (const x of nums) {
        f = Math.max(f, 0) + x;
        g = Math.min(g, 0) + x;
        ans = Math.max(ans, f, -g);
    }
    return ans;
}

Rust Code
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13

impl Solution {
    pub fn max_absolute_sum(nums: Vec<i32>) -> i32 {
        let mut f = 0;
        let mut g = 0;
        let mut ans = 0;
        for x in nums {
            f = i32::max(f, 0) + x;
            g = i32::min(g, 0) + x;
            ans = i32::max(ans, f.max(-g));
        }
        ans
    }
}