14. Fixing Two nodes of a BST

The problem can be found at the following link: Question Link

Problem Description

Given the root of a Binary Search Tree (BST), where exactly two nodes were swapped by mistake, your task is to fix (or correct) the BST by swapping them back. The structure of the tree should not change.

Examples

Example 1:

Input:

        10
       /  \
      5    8
     /  \
    2    20

Output:

1

Explanation:

The nodes 20 and 8 were swapped by mistake. After swapping them back, the BST is restored correctly.

Example 2:

Input:

        5
       / \
     10   20
     / \
    2   8

Output:

1

Explanation:

The nodes 10 and 5 were swapped by mistake. After swapping them back, the BST is restored correctly.

Constraints:

• $(1 \leq \text{Number of Nodes} \leq 10^3)$

My Approach

Optimized Inorder Traversal (O(N) Time, O(H) Space)

1. Use an inorder traversal to detect swapped nodes in the BST.

2. Identify the two misplaced nodes:

   • If a node appears larger than the next node, it's incorrectly placed.

   • Track the first misplaced node and the second misplaced node.

3. Swap the values of the two misplaced nodes to restore the BST.

Algorithm Steps:

1. Perform an inorder traversal to find the two misplaced nodes.

2. If the first misplaced node is found, store it in first.

3. If a second misplaced node is found later, store it in last.

4. If there's no second misplaced node, use the middle node instead.

5. Swap the values of the two misplaced nodes.

Time and Auxiliary Space Complexity

• Expected Time Complexity: O(N), since we traverse each node once.

• Expected Auxiliary Space Complexity: O(H), due to the recursion stack in the inorder traversal.

Code (C++)

class Solution {
public:
    void correctBST(Node* root) {
        Node *first = nullptr, *middle = nullptr, *last = nullptr, *prev = nullptr;
        function<void(Node*)> inorder = [&](Node* node) {
            if (!node) return;
            inorder(node->left);
            if (prev && node->data < prev->data) {
                if (!first) first = prev, middle = node;
                else last = node;
            }
            prev = node;
            inorder(node->right);
        };
        inorder(root);
        swap(first->data, last ? last->data : middle->data);
    }
};

🌲 Alternative Approaches

2️⃣ Iterative Inorder Traversal (Stack)

Approach

1. Use a stack for inorder traversal (instead of recursion).

2. Detect swapped nodes by checking the inorder order.

3. Swap the incorrect nodes back to restore the BST.

class Solution {
public:
    void correctBST(Node* root) {
        stack<Node*> st;
        Node *first = nullptr, *middle = nullptr, *last = nullptr, *prev = nullptr;

        while (!st.empty() || root) {
            while (root) {
                st.push(root);
                root = root->left;
            }
            root = st.top(); st.pop();
            if (prev && root->data < prev->data) {
                if (!first) first = prev, middle = root;
                else last = root;
            }
            prev = root;
            root = root->right;
        }

        swap(first->data, last ? last->data : middle->data);
    }
};

🔹 Avoids recursion stack overflow issues using an explicit stack.

3️⃣ Morris Traversal (O(1) Space)

Approach

1. Use Morris Traversal to perform an inorder traversal without extra space.

2. Identify misplaced nodes while modifying the BST structure temporarily.

3. Restore the BST by swapping the misplaced nodes.

class Solution {
public:
    void correctBST(Node* root) {
        Node *first = nullptr, *middle = nullptr, *last = nullptr, *prev = nullptr;

        while (root) {
            if (!root->left) {
                if (prev && root->data < prev->data) {
                    if (!first) first = prev, middle = root;
                    else last = root;
                }
                prev = root;
                root = root->right;
            } else {
                Node* pre = root->left;
                while (pre->right && pre->right != root) pre = pre->right;
                if (!pre->right) {
                    pre->right = root;
                    root = root->left;
                } else {
                    pre->right = nullptr;
                    if (prev && root->data < prev->data) {
                        if (!first) first = prev, middle = root;
                        else last = root;
                    }
                    prev = root;
                    root = root->right;
                }
            }
        }
        swap(first->data, last ? last->data : middle->data);
    }
};

🔹 Uses O(1) space without recursion or extra stack.

Comparison of Approaches

Approach	⏱️ Time Complexity	🗂️ Space Complexity	⚡ Method	✅ Pros	⚠️ Cons
Recursive Inorder	🟢 O(N)	🟡 O(H)	Recursion	Simple and easy to implement	Uses recursion stack space
Iterative Inorder	🟢 O(N)	🟡 O(H)	Stack-based	Avoids recursion depth issues	Uses extra memory for stack
Morris Traversal	🟢 O(N)	🟢 O(1)	No extra space	No additional memory needed	Modifies tree temporarily

💡 Best Choice?

• ✅ For space efficiency: Morris Traversal (O(1) space).

• ✅ For simplicity: Recursive Inorder Traversal.

• ✅ For large trees: Iterative Inorder Traversal avoids recursion depth issues.

Code (Java)

class Solution {
    Node first, middle, last, prev;

    void inorder(Node root) {
        if (root == null) return;
        inorder(root.left);
        if (prev != null && root.data < prev.data) {
            if (first == null) {
                first = prev;
                middle = root;
            } else {
                last = root;
            }
        }
        prev = root;
        inorder(root.right);
    }

    void correctBST(Node root) {
        first = middle = last = prev = null;
        inorder(root);
        int temp = first.data;
        first.data = (last != null) ? last.data : middle.data;
        if (last != null) last.data = temp;
        else middle.data = temp;
    }
}

Code (Python)

class Solution:
    def correctBST(self, root):
        self.first = self.middle = self.last = self.prev = None

        def inorder(node):
            if not node:
                return
            inorder(node.left)
            if self.prev and node.data < self.prev.data:
                if not self.first:
                    self.first, self.middle = self.prev, node
                else:
                    self.last = node
            self.prev = node
            inorder(node.right)

        inorder(root)
        self.first.data, (self.last or self.middle).data = (self.last or self.middle).data, self.first.data

Contribution and Support

For discussions, questions, or doubts related to this solution, feel free to connect on LinkedIn: Any Questions. Let’s make this learning journey more collaborative!

⭐ If you find this helpful, please give this repository a star! ⭐

---

📍Visitor Count