> For the complete documentation index, see [llms.txt](https://aaronice.gitbook.io/lintcode/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://aaronice.gitbook.io/lintcode/array/maximum-subarray-difference.md).

# Maximum Subarray Difference

### Description

Given an array with integers.

Find two*non-overlapping\_subarrays\_A\_and\_B*, which`|SUM(A) - SUM(B)|`is the largest.

Return the largest difference.

> The subarray should contain at least one number

### Example

For`[1, 2, -3, 1]`, return`6`.

### Challenge

O(n) time and O(n) space.

## Analysis

与Maximum Subarray系列问题一样，这里其实只用转化问题即，求 max |SUM(A) - SUM(B)|，等价为求

`Math.max(ans, Math.max(Math.abs(leftMax(A) - rightMin(B)), Math.abs(leftMin(A) - rightMax(B))))`

也就是和之前求 maximum non-overlap subarrays 的思路基本一致，只不过要求的数组变成了四个：`leftMax`, `leftMin`, `rightMax` 和 `rightMin`

虽然最后代码很长，但是原理很简单，总共5 pass，4次pass用来生成`leftMax`, `leftMin`, `rightMax` 和 `rightMin`

最后一个pass用来寻找最终的答案。因此时间复杂度O(n), 空间复杂度O(n).

## Solution

Kadane's Algorithm - O(n) space, O(n) time (283ms beats 96.40%!)

```java
public class Solution {
    /**
     * @param nums: A list of integers
     * @return: An integer indicate the value of maximum difference between two substrings
     */
    public int maxDiffSubArrays(int[] nums) {
        if (nums == null || nums.length == 0) return 0;
        int n = nums.length;
        int[] leftMax = new int[n];
        int[] leftMin = new int[n];
        int[] rightMax = new int[n];
        int[] rightMin = new int[n];

        leftMax[0] = nums[0];
        leftMin[0] = nums[0];
        rightMax[n - 1] = nums[n - 1];
        rightMin[n - 1] = nums[n - 1];

        int localMax = leftMax[0];
        int globalMax = leftMax[0];
        for (int i = 1; i < n; i++) {
            localMax = Math.max(localMax + nums[i], nums[i]);
            globalMax = Math.max(localMax, globalMax);
            leftMax[i] = globalMax;
        }

        int localMin = leftMin[0];
        int globalMin = leftMin[0];
        for (int i = 1; i < n; i++) {
            localMin = Math.min(localMin + nums[i], nums[i]);
            globalMin = Math.min(localMin, globalMin);
            leftMin[i] = globalMin;
        }

        localMax = rightMax[n - 1];
        globalMax = rightMax[n - 1];
        for (int i = n - 2; i >= 0; i--) {
            localMax = Math.max(localMax + nums[i], nums[i]);
            globalMax = Math.max(localMax, globalMax);
            rightMax[i] = globalMax;
        }

        localMin = rightMin[n - 1];
        globalMin = rightMin[n - 1];
        for (int i = n - 2; i >= 0; i--) {
            localMin = Math.min(localMin + nums[i], nums[i]);
            globalMin = Math.min(localMin, globalMin);
            rightMin[i] = globalMin;
        }

        int result = Integer.MIN_VALUE;
        for (int i = 0; i < n - 1; i++) {
            int temp = Math.max(Math.abs(leftMax[i] - rightMin[i + 1]), Math.abs(leftMin[i] - rightMax[i + 1]));
            result = Math.max(temp, result);
        }

        return result;
    }
}
```


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://aaronice.gitbook.io/lintcode/array/maximum-subarray-difference.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.