# Count subarrays where second highest lie before highest

Given an array of N distinct element of at least size 2. A pair (a, b) in an array is defined as ‘a’ is the index of second smallest element and ‘b’ is the index of highest element in the array. The task is to count all the distinct pair where a < b in all the subarrays.

Examples :

```Input : arr[] = { 1, 3, 2, 4 }
Output : 3
```

Explanation :
The subarray { 1 }, { 3 }, { 2 }, { 4 } does not contain any such pair
The subarray { 1, 3 }, { 2, 4 } contain (1, 2) as pair
The subarray { 3, 2 } does not contain any such pair
The subarray { 3, 2, 4 } contain (1, 3) as a pair
The subarray { 1, 3, 2 } does not contain any such pair
The subarray { 1, 3, 2, 4 } contain (2, 4) as a pair
So, there are 3 distinct pairs, which are (1, 2), (1, 3) and (2, 4).

## Recommended: Please try your approach on {IDE} first, before moving on to the solution.

Method 1: (Brute Force) – Simple approach can be,
1. Find all the subarrays.
2. For each subarray, find second largest and largest element.
3. Check if second largest element lie before largest element.
4. If so, check if such index pair is already counted or not. If not store the pair and increment the count by 1, else ignore.
Time Complexity: O(n2).

Method 2: (Using stack) –
For given array A, suppose for an element at index curr (A[curr]), first element greater than it and after it is A[next] and first element greater than it and before it A[prev]. Observe that for all subarray starting from any index in range [prev + 1, curr] and ending at index next, A[curr] is the second largest and A[next] is the largest, which generate (curr – prev + 1) pairs in total with difference of (next – curr + 1) in maximum and second maximum.

If we get next and prev greater element in an array, and keep track of maximum number of pairs possible for any difference (of largest and second largest). We will need to add all these numbers.

Now only job left is to get greater element (after and before) any element. For this, refer Next Greater Element.

Traverse from the starting element in an array, keeping track of all numbers in the stack in decreasing order. After arriving at any number, pop all elements from stack which are less than current element to get location of number bigger than it and push current element on it. This generates required value for all numbers in the array.

 `// C++ program to count number of distinct instance ` `// where second highest number lie ` `// before highest number in all subarrays. ` `#include ` `#define MAXN 100005 ` `using` `namespace` `std; ` ` `  `// Finding the next greater element of the array. ` `void` `makeNext(``int` `arr[], ``int` `n, ``int` `nextBig[]) ` `{ ` `    ``stack > s; ` ` `  `    ``for` `(``int` `i = n - 1; i >= 0; i--) { ` ` `  `        ``nextBig[i] = i; ` `        ``while` `(!s.empty() && s.top().first < arr[i]) ` `            ``s.pop(); ` ` `  `        ``if` `(!s.empty()) ` `            ``nextBig[i] = s.top().second; ` ` `  `        ``s.push(pair<``int``, ``int``>(arr[i], i)); ` `    ``} ` `} ` ` `  `// Finding the previous greater element of the array. ` `void` `makePrev(``int` `arr[], ``int` `n, ``int` `prevBig[]) ` `{ ` `    ``stack > s; ` `    ``for` `(``int` `i = 0; i < n; i++) { ` ` `  `        ``prevBig[i] = -1; ` `        ``while` `(!s.empty() && s.top().first < arr[i]) ` `            ``s.pop(); ` ` `  `        ``if` `(!s.empty()) ` `            ``prevBig[i] = s.top().second; ` ` `  `        ``s.push(pair<``int``, ``int``>(arr[i], i)); ` `    ``} ` `} ` ` `  `// Wrapper Function ` `int` `wrapper(``int` `arr[], ``int` `n) ` `{ ` `    ``int` `nextBig[MAXN]; ` `    ``int` `prevBig[MAXN]; ` `    ``int` `maxi[MAXN]; ` `    ``int` `ans = 0; ` ` `  `    ``// Finding previous largest element ` `    ``makePrev(arr, n, prevBig); ` ` `  `    ``// Finding next largest element ` `    ``makeNext(arr, n, nextBig); ` ` `  `    ``for` `(``int` `i = 0; i < n; i++) ` `        ``if` `(nextBig[i] != i) ` `            ``maxi[nextBig[i] - i] = max(maxi[nextBig[i] - i], ` `                                       ``i - prevBig[i]); ` ` `  `    ``for` `(``int` `i = 0; i < n; i++) ` `        ``ans += maxi[i]; ` ` `  `    ``return` `ans; ` `} ` ` `  `// Driven Program ` `int` `main() ` `{ ` `    ``int` `arr[] = { 1, 3, 2, 4 }; ` `    ``int` `n = ``sizeof``(arr) / ``sizeof``(arr[0]); ` `    ``cout << wrapper(arr, n) << endl; ` `    ``return` `0; ` `} `

Output :

```3
```

Time Complexity: O(n)

Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above.

This article is attributed to GeeksforGeeks.org

## tags:

Arrays Stack cpp-pair cpp-stack Arrays Stack

code

load comments