Related
I am trying to print the kth smallest element in an BST.
The first solution is using in-order traversal.
Next solution is finding the index of the current node by calculation the size of its left subtree.
Complete algo:
Find size of left subtree:
1.If size = k-1, return current node
2.If size>k return (size-k)th node in right subtree
3.If size<k return kth node in left subtree
This can be implemented using a separate count function which looks something like
public class Solution {
public int kthSmallest(TreeNode root, int k) {
//what happens if root == null
//what happens if k > total size of tree
return kthSmallestNode(root,k).val;
}
public static TreeNode kthSmallestNode(TreeNode root,int k){
if(root==null) return root;
int numberOfNodes = countNodes(root.left);
if(k == numberOfNodes ) return root;
if(k<numberOfNodes ) return kthSmallestNode(root.left,k);
else return kthSmallestNode(root.right,k-numberOfNodes );
}
private static int countNodes(TreeNode node){
if(node == null) return 0;
else return 1+countNodes(node.left)+countNodes(node.right);
}
}
But I see that we count the size for same trees multiple times, so one way is to maintain an array to store thes sizes like the DP way.
But I want to write a recursive solution for this.And here is the code I have written.
class Node {
int data;
Node left;
Node right;
public Node(int data, Node left, Node right) {
this.left = left;
this.data = data;
this.right = right;
}
}
public class KthInBST
{
public static Node createBST(int headData)
{
Node head = new Node(headData, null, null);
//System.out.println(head.data);
return head;
}
public static void insertIntoBst(Node head, int data)
{
Node newNode = new Node(data, null, null);
while(true) {
if (data > head.data) {
if (head.right == null) {
head.right = newNode;
break;
} else {
head = head.right;
}
} else {
if (head.left == null) {
head.left = newNode;
break;
} else {
head = head.left;
}
}
}
}
public static void main(String[] args)
{
Node head = createBST(5);
insertIntoBst(head, 7);
insertIntoBst(head, 6);
insertIntoBst(head, 2);
insertIntoBst(head, 1);
insertIntoBst(head, 21);
insertIntoBst(head, 11);
insertIntoBst(head, 14);
insertIntoBst(head, 3);
printKthElement(head, 3);
}
public static int printKthElement(Node head, int k)
{
if (head == null) {
return 0;
}
int leftIndex = printKthElement(head.left, k);
int index = leftIndex + 1;
if (index == k) {
System.out.println(head.data);
} else if (k > index) {
k = k - index;
printKthElement(head.right, k);
} else {
printKthElement(head.left, k);
}
return index;
}
}
This is printing the right answer but multiple times, I figured out why it is printing multiple times but not understanding how to avoid it.
And also If I want to return the node instead of just printing How do I do it?
Can anyone please help me with this?
Objective:
Recursively finding the kth smallest element in a binary search tree and returning the node corresponding to that element.
Observation:
The number of elements smaller than the current element is the size of the left subtree so instead of recursively calculating its size, we introduce a new member in class Node, that is, lsize which represents the size of the left subtree of current node.
Solution:
At each node we compare the size of left subtree with the current value of k:
if head.lsize + 1 == k: current node in our answer.
if head.lsize + 1 > k: elements in left subtree are more than k, that is, the k the smallest element lies in the left subtree. So, we go left.
if head.lsize + 1 < k: the current element alongwith all the elements in the left subtree are less than the kth element we need to find. So, we go to the right subtree but also reduce k by the amount of elements in left subtree + 1(current element). By subtracting this from k we make sure that we have already taken into account the number of elements which are smaller than k and are rooted as the left subtree of current node (including the current node itself).
Code:
class Node {
int data;
Node left;
Node right;
int lsize;
public Node(int data, Node left, Node right) {
this.left = left;
this.data = data;
this.right = right;
lsize = 0;
}
}
public static void insertIntoBst(Node head, int data) {
Node newNode = new Node(data, null, null);
while (true) {
if (data > head.data) {
if (head.right == null) {
head.right = newNode;
break;
} else {
head = head.right;
}
} else {
head.lsize++; //as we go left, size of left subtree rooted
//at current node will increase, hence the increment.
if (head.left == null) {
head.left = newNode;
break;
} else {
head = head.left;
}
}
}
}
public static Node printKthElement(Node head, int k) {
if (head == null) {
return null;
}
if (head.lsize + 1 == k) return head;
else if (head.lsize + 1 > k) return printKthElement(head.left, k);
return printKthElement(head.right, k - head.lsize - 1);
}
Changes:
A new member lsize has been introduced in class Node.
Slight modification in insertIntoBst.
Major changes in printKthElement.
Corner case:
Add a check to ensure that k is between 1 and the size of the tree otherwise a null node will be returned resulting in NullPointerException.
This is working on the test cases I have tried, so far. Any suggestions or corrections are most welcome.
:)
I want to find the nth node/element in a binary tree. Not the nth largest/smallest, just the nth in inorder order for example.
How would this be done? Is it possible to keep it to one function? Many functions employ an external variable to keep track of the iterations outside of the recursion, but that seems... lazy, for lack of a better term.
You can augment the binary search tree into an order statistic tree, which supports a "return the nth element" operation
Edit: If you just want the ith element of an inorder traversal (instead of the ith smallest element) and don't want to use external variables then you can do something like the following:
class Node {
Node left
Node right
int data
}
class IterationData {
int returnVal
int iterationCount
}
IterationData findNth(Node node, IterationData data, int n) {
if(node.left != null) {
data = findNth(node.left, data, n)
}
if(data.iterationCount < n) {
data.iterationCount++
if(data.iterationCount == n) {
data.returnVal = node.data
return data
} else if(node.right != null) {
return findNth(node.right, data, n)
} else {
return data
}
}
}
You'll need some way to return two values, one for the iteration count and one for the return value once the nth node is found; I've used a class, but if your tree contains integers then you could use an integer array with two elements instead.
In order iterative traversal, keep track of nodes passed in external variable.
public static Node inOrderInterativeGet(Node root, int which){
Stack<Node> stack = new Stack<Node>();
Node current = root;
boolean done = false;
int i = 1;
if(which <= 0){
return null;
}
while(!done){
if(current != null){
stack.push(current);
current = current.getLeft();
}
else{
if(stack.empty()){
done = true;
}
else{
current = stack.pop();
if(i == which){
return current;
}
i++;
current = current.getRight();
}
}
}
return null;
}
One way to do it is to have a size property which is left_subtree + right_subtree + 1:
class Node:
def __init__(self, data=None, left=None, right=None,
size=None):
self.data = data
self.left = left
self.right = right
self.size = size
def select(node, k):
"""Returns node with k-th position in-order traversal."""
if not node:
return None
t = node.left.size if node.left else 0
if t > k:
return select(node.left, k)
elif t < k:
return select(node.right, k - t - 1)
else:
return node
If you don't like global variable, pass to recursive function additional parameter - some int variable, let's call it auto_increment or just ai. ai stores order of current node. Also, recursive function should return maximal value of ai of current vertex subtree,because after visiting whole subtree next 'free' value will be max_ai_in_subreee+1 Something like that
int rec(int vertex,int ai){
traverseOrder[vertex] = ai
if(getChildren(vertex)!=null) return ai;
else{
for(childrens){
ai = rec(child,ai+1);
}
return ai;// subtree visited, return maximal free value upstairs.
}
}
If your function already returns some useful data, it may return some complex object which contains {useful data+ai}
Start from some vertex looks like rec(start_vertex,1);
Below is full code that you can use to find the nth element using inorder in a Binary Tree.
public class NthNodeInInoeder {
static public class Tree {
public int data;
public Tree left,right;
public Tree(int data) {
this.data = data;
}
}
static Tree root;
static int count = 0;
private static void inorder(Tree root2, int num) {
if (root2 == null)
return;
Tree node = root2;
Stack<Tree> stack = new Stack<>();
while (node != null || stack.size() > 0) {
while (node != null) {
stack.push(node);
node = node.left;
}
node = stack.pop();
count++;
if (count == num) {
System.out.println(node.data);
break;
}
node = node.right;
}
}
public static void main(String[] args) {
Scanner sc = new Scanner(System.in);
root = new Tree(10);
root.left = new Tree(20);
root.right = new Tree(30);
root.left.left = new Tree(40);
root.left.right = new Tree(50);
int num = sc.nextInt();
inorder(root, num);
sc.close();
}
}
I need to implement a non-recursive function to determine if a binary tree is balanced or not.
Anyone?
Thanks!!!
Assuming that by "balanced", you mean "height-balanced" in the AVL-tree sense, and you can store arbitrary information for each node,
For each node in post-order,
if either child doesn't exist, assume its respective height is 0.
if the height of both children differs by more than one, the tree is not balanced.
otherwise, this node's height is the larger of both children's heights.
If this point is reached, the tree is balanced.
One way to perform post-order traversal:
start at the root
loop
if this node's left child exists and does not have its height computed, visit its left child next.
else if this node's right child exists and does not have its height computed, visit its right child next.
else
compute this node's height, possibly returning early
if this node is not the root, visit its parent next.
If this point is reached, the tree is balanced.
Try this,
public class BalancedBinaryTree {
public static class TreeNode {
int val;
TreeNode left;
TreeNode right;
TreeNode() {}
TreeNode(int val) { this.val = val; }
TreeNode(int val, TreeNode left, TreeNode right) {
this.val = val;
this.left = left;
this.right = right;
}
}
public boolean isBalanced(TreeNode root) {
if (root==null) {
return true;
}
Stack<TreeNode> stack = new Stack<>();
Map<TreeNode, Integer> map = new HashMap<>();
stack.push(root);
while(!stack.isEmpty()) {
TreeNode node = stack.pop();
if ((node.left==null || (node.left!=null && map.containsKey(node.left))) && (node.right==null || (node.right!=null && map.containsKey(node.right)))) {
int right = (node.right==null) ? 0 : map.get(node.right);
int left = (node.left==null) ? 0 : map.get(node.left);
if (Math.abs(right-left)>1) {
return false;
} else {
map.put(node, Math.max(right, left)+1);
}
} else {
if (node.left!=null && !map.containsKey(node.left)) {
stack.push(node);
stack.push(node.left);
} else {
stack.push(node);
stack.push(node.right);
}
}
}
return true;
}
public static void main(String[] args) {
BalancedBinaryTree b = new BalancedBinaryTree();
boolean v = b.isBalanced(new TreeNode(3, new TreeNode(9), new TreeNode(20, new TreeNode(15), new TreeNode(7))));
System.out.println(v);
v = b.isBalanced(new TreeNode(1, new TreeNode(2, new TreeNode(3, new TreeNode(4), new TreeNode(4)), new TreeNode(3)), new TreeNode(2)));
System.out.println(v);
v = b.isBalanced(new TreeNode(1, new TreeNode(2, new TreeNode(4, new TreeNode(8), null), new TreeNode(5)), new TreeNode(3, new TreeNode(6), null)));
System.out.println(v);
}
}
Here is a c++ code that works, inspired by the postorder traversal. The code is not commented because i do not think a simple comment is enough to explain the whole algorithm. You can execute this code manually with the example below and then you will understand everything.
bool IsBalance(const Node *head)
{
std::stack<const Node *> s;
std::stack<int> sV;
const Node *curr = head, *lastVisit = nullptr;
int deep = 0;
while (curr || !s.empty())
{
while (curr)
{
s.push(curr);
sV.push(-1);
curr = curr->m_pLeft;
}
curr = s.top();
if (sV.top() == -1)
{
sV.top() = deep;
}
if (!curr->m_pRight || curr->m_pRight == lastVisit)
{
if (!curr->m_pRight)
{
deep = 0;
}
if (std::abs(sV.top() - deep) > 1)
{
return false;
}
deep = std::max(sV.top(), deep) + 1;
lastVisit = curr;
s.pop();
sV.pop();
curr = nullptr;
}
else
{
deep = 0;
curr = curr->m_pRight;
}
}
return true;
}
examples:
(1) 21,10,3,1,#,#,5,#,6,#,#,15,12,#,#,18,16,#,#,20,#,#,35,30,22,#,#,#,40,36,#,#,42,#,45,#,#
(2) 1,2,#,4,#,5,#,#,3,6,8,#,#,#,7,#,#
(3) 3,1,#,2,#,#,#
Where nodes are arranged by PreOrder, separated by commas, and # indicates an empty node.
Find the height of left subtree and right subtree for a node of the tree, using Level order traversal and check if that node is balanced.
Repeat this for every node of the tree. For traversing all the nodes we can use level order traversal to avoid recursion.
int height(TreeNode* root){
if(!root){
return 0;
}
queue<TreeNode*> q;
q.push(root);
int count=0;
while(!q.empty()){
int size=q.size();
for(int i=0;i<size;++i){
TreeNode* temp=q.front();
q.pop();
if(temp->left){
q.push(temp->left);
}
if(temp->right){
q.push(temp->right);
}
}
count++;
}
return count;
}
bool checkEveryNode(TreeNode* root){
if(!root){
return true;
}
queue<TreeNode*> q;
q.push(root);
while(!q.empty()){
int count=q.size();
for(int i=0;i<count;++i){
TreeNode* temp=q.front();
q.pop();
int left=height(temp->left);
int right=height(temp->right);
if(abs(left-right)>1){
return false;
}
if(temp->left){
q.push(temp->left);
}
if(temp->right){
q.push(temp->right);
}
}
}
return true;
}
bool isBalanced(TreeNode* root) {
return checkEveryNode(root);
}
Time complexity of this approach is O(n^2), as we need to traverse all the descendant nodes for finding the height of a node(N) and we need to do this for all the nodes(N)
I was reading through Introduction to algorithms i came across this problem about In-order Traversal of binary search tree without using a stack or recursion. Hint says to assume that testing of pointers for equality is a legitimate operation.I'm stuck finding the solution to this problem. Please give me some direction. I'm not looking for the code. Just give me the right direction.
Exact duplicate here
No stack nor recursion means you have to use pointers. Not giving you code nor the exact answer, since you asked not to :)
Think about how to explore the tree without using recursion: what would you need to do? What pointer(s) do you need to keep? Can a tree node have a pointer to the parent?
Hope it helps.
We need a Threaded Binary Tree to do in-order Traversal without recursion / stack.
Wiki says 'A binary tree is threaded by making all right child pointers that would normally be null point to the inorder successor of the node, and all left child pointers that would normally be null point to the inorder predecessor of the node'
So you are given a normal Binary Tree , Convert it into a Threaded Binary Tree which can be done using Morris Traversal.
What you are going to do in Morris Traversal is to connect each node with its in-order successor. So while visiting a node ,Search for its in-order predecessor and let it be Pred.
then make Pred->right=Current node and we have to revert back the changes too. You can better refer this http://www.geeksforgeeks.org/archives/6358 for a great explanation.
Hello Parminder i have implemented your question in java.Please check it once
class InorderWithoutRecursion {
public static void morrisTraversal(TreeNode root) {
TreeNode current,pre;
if(root == null)
return;
current = root;
while(current != null){
if(current.left == null){
System.out.println(current.data);
current = current.right;
}
else {
/* Find the inorder predecessor of current */
pre = current.left;
while(pre.right != null && pre.right != current)
pre = pre.right;
/* Make current as right child of its inorder predecessor */
if(pre.right == null){
pre.right = current;
current = current.left;
}
/* Revert the changes made in if part to restore the original
tree i.e., fix the right child of predecssor */
else {
pre.right = null;
System.out.println(current.data);
current = current.right;
}
}
}
}
public static void main(String[] args) {
int[] nodes_flattened = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10};
TreeNode root = TreeNode.createMinimalBST(nodes_flattened);
morrisTraversal(root);
}
}
For TreeNode class below code will help you..
public class TreeNode {
public int data;
public TreeNode left;
public TreeNode right;
public TreeNode parent;
public TreeNode(int d) {
data = d;
}
public void setLeftChild(TreeNode left) {
this.left = left;
if (left != null) {
left.parent = this;
}
}
public void setRightChild(TreeNode right) {
this.right = right;
if (right != null) {
right.parent = this;
}
}
public void insertInOrder(int d) {
if (d <= data) {
if (left == null) {
setLeftChild(new TreeNode(d));
} else {
left.insertInOrder(d);
}
} else {
if (right == null) {
setRightChild(new TreeNode(d));
} else {
right.insertInOrder(d);
}
}
}
public boolean isBST() {
if (left != null) {
if (data < left.data || !left.isBST()) {
return false;
}
}
if (right != null) {
if (data >= right.data || !right.isBST()) {
return false;
}
}
return true;
}
public int height() {
int leftHeight = left != null ? left.height() : 0;
int rightHeight = right != null ? right.height() : 0;
return 1 + Math.max(leftHeight, rightHeight);
}
public TreeNode find(int d) {
if (d == data) {
return this;
} else if (d <= data) {
return left != null ? left.find(d) : null;
} else if (d > data) {
return right != null ? right.find(d) : null;
}
return null;
}
private static TreeNode createMinimalBST(int arr[], int start, int end){
if (end < start) {
return null;
}
int mid = (start + end) / 2;
TreeNode n = new TreeNode(arr[mid]);
n.setLeftChild(createMinimalBST(arr, start, mid - 1));
n.setRightChild(createMinimalBST(arr, mid + 1, end));
return n;
}
public static TreeNode createMinimalBST(int array[]) {
return createMinimalBST(array, 0, array.length - 1);
}
}
This is an Amazon interview question. Can any one give an algorithm to do this?
There is a binary tree with the following properties:
All of its inner node have the value 'N', and all the leaves have the value 'L'.
Every node either has two children or has no child.
Given its preorder, construct the tree and return the root node.
Since it is guaranteed that each internal node has exactly 2 children, we can simply build the tree recursively using that.
We call our function with the input provided, and it examines the first character it got. If it is a leaf node, it just returns a leaf. If it is an internal node, it just calls itself for the left and right subtrees and returns the tree formed using the node as root and the left and right subtrees as its left and right children.
Code follows (in Python). Note, I am using tuples to represent node, so the tree is a tuple of tuples.
#! /usr/bin/env python
from collections import deque
def build_tree(pre_order):
root=pre_order.popleft()
if root=='L':
return root
else:
return (root,build_tree(pre_order),build_tree(pre_order))
if __name__=='__main__':
print build_tree(deque("NNLLL"))
Edit: Code in Java
import java.util.*;
class Preorder{
public static Node buildTree(List<Character> preorder){
char token=preorder.remove(0);
if (token=='L'){
return new Node(token,null,null);
}
else{
return new Node(token,buildTree(preorder),buildTree(preorder));
}
}
public static void main(String args[]){
List<Character> tokens=new LinkedList<Character>();
String input="NNLLL";
for(int i=0;i<input.length();i++) tokens.add(input.charAt(i));
System.out.println(buildTree(tokens));
}
}
class Node{
char value;
Node left,right;
public Node(char value, Node left, Node right){
this.value=value;
this.left=left;
this.right=right;
}
public String toString(){
if (left==null && right==null){
return "("+value+")";
}
else{
return "("+value+", "+left+", "+right+")";
}
}
}
I can think of a recursive algorithm.
head = new node.
remove first character in preorderString
Invoke f(head, preorderString)
Recursive function f(node, s)
- remove first char from s, if L then attach to node as leaf.
else create a nodeLeft, attach to node, invoke f(nodeLeft, s)
- remove first char from s, if L then attach to node as leaf.
else create a nodeRight, attach to node, invoke f(nodeRight, s)
I think the key point is to realize that there are three possibilities for the adjacent nodes: NN, NL?, L? (``?'' means either N or L)
NN: the second N is the left child of the first N, but we don't know what the right child of the first N is
NL?: the second N is the left child of the first N, and the right child of the first N is ?
L?: ? is the right child of STACK top
A STACK is used because when we read a node in a preorder sequence, we don't know where its right child is (we do know where its left child is, as long as it has one). A STACK stores this node so that when its right child appears we can pop it up and finish its right link.
NODE * preorder2tree(void)
{
NODE * head = next_node();
NODE * p = head;
NODE * q;
while (1) {
q = next_node();
if (!q)
break;
/* possibilities of adjacent nodes:
* NN, NL?, L?
*/
if (p->val == 'N') {
p->L = q;
if (q->val == 'N') { /* NN */
push(p);
p = q;
} else { /* NL? */
q = next_node();
p->R = q;
p = q;
}
} else { /* L? */
p = pop();
p->R = q;
p = q;
}
}
return head;
}
The code above was tested using some simple cases. Hopefully it's correct.
Here is the java program::
import java.util.*;
class preorder_given_NNNLL
{
static Stack<node> stk = new Stack<node>();
static node root=null;
static class node
{
char value;
node left;
node right;
public node(char value)
{
this.value=value;
this.left=null;
this.right=null;
}
}
public static node stkoper()
{
node posr=null,posn=null,posl=null;
posr=stk.pop();
if(stk.empty())
{
stk.push(posr);
return null;
}
else
posl=stk.pop();
if(stk.empty())
{
stk.push(posl);
stk.push(posr);
return null;
}
else
{
posn=stk.pop();
}
if( posn.value == 'N' && posl.value == 'L' && posr.value == 'L')
{
root = buildtree(posn, posl, posr);
if(stk.empty())
{
return root;
}
else
{
stk.push(root);
root=stkoper();
}
}
else
{
stk.push(posn);
stk.push(posl);
stk.push(posr);
}
return root;
}
public static node buildtree(node posn,node posl,node posr)
{
posn.left=posl;
posn.right=posr;
posn.value='L';
return posn;
}
public static void inorder(node root)
{
if(root!=null)
{
inorder(root.left);
if((root.left == null) && (root.right == null))
System.out.println("L");
else
System.out.println("N");
inorder(root.right);
}
}
public static void main(String args[]){
String input="NNNLLLNLL";
char[] pre = input.toCharArray();
for (int i = 0; i < pre.length; i++)
{
node temp = new node(pre[i]);
stk.push(temp);
root=stkoper();
}
inorder(root);
}
}
The construct function does the actual tree construction. The code snippet is the solution for the GeeksforGeeks question that you mentioned as above.
struct Node*construct(int &index, Node*root, int pre[], int n, char preLN[])
{
Node*nodeptr;
if(index==n)
{
return NULL;
}
if(root==NULL)
{
nodeptr = newNode(pre[index]);
}
if(preLN[index]=='N')
{
index = index+1;
nodeptr->left = construct(index, nodeptr->left, pre,n,preLN);
index = index+1;
nodeptr->right = construct(index, nodeptr->right,pre,n,preLN);
return nodeptr;
}
return nodeptr;
}
struct Node *constructTree(int n, int pre[], char preLN[])
{
int index =0;
Node*root = construct(index,NULL,pre,n,preLN);
return root;
}
Points to Note:
Index has been declared a reference variable so that on returning back to the parent node, the function starts constructing the tree from the overall most recent value of index and not the value of index as possessed by the function when initially executing the call.
Different values of index for right and left subtrees since preorder traversal follows Root, Left ,Right sequence of nodes.
Hope it Helps.