Skip to main content

Binary tree deletion

Do not get all scared and worried. It is not rocket science (or as I would like to call it - it is not regex).

Remember deleting a node from linked list. When you deleted a node, you did the following 2 steps
  1. Free the memory of the node
  2. Link the previous node of the deleted node to the next node

You will have to do these things in binary tree too. But the difficulty here is do you link the previous node - or parent node in tree terminology, to the left child of deleted node? Or to the right child? You can not link both children, as the parent already may have one child node.

So before we solve this, let us categorize the deletion with the help of a diagram.



Consider the cases

  1. The node to be deleted is a leaf node. That is, it does not have left or right child. In this case, link the parent to NULL in place of deleted node. 
    • If you want to delete 13 - which is a leaf node, you set 14->left to NULL and free memory of node 13.
    • If you want to delete 7 - which is also a leaf node and is the right child of 6, you set 6->right to NULL and free 6.
  2. The node is to deleted is having only one subtree and hence one child node. In this case link the parent to this child node.
    • If you want to delete 14 - which has only left child node, you link the only child of 14 to the parent. That is you set 10->right to 14->left. Now 10 will have right child as 13. And BST property holds good.
    • Instead if you want to delete 10 which has only right subtree - 14 and 13, you link parent of 10 to  14. That is, you set 8->right to 14. 
  3. The node to be deleted has both branches. In this case, you find the successor of the node - which will be treeminimum of right sub tree. Copy the data of this tree minimum to the node to be deleted. And finally delete this tree minimum.
    • If you want to delete 3, you search for minimum of right subtree of 3. Right subtree of 3 has 3 nodes in it, 6,4 and 7. Minimum of this is 4. Copy the data - 4 into the node to be deleted -which is 3. Now you delete the node with 4.  
Delete the node with only left subtree

Delete the node with only right subtree

Delete the node with both subtrees

 Why tree minimum of right subtree?

When you delete a node, its place can be taken by its predecessor or its successor. We are choosing successor here is a substitute for node to be deleted. And we are preserving its information by copying it into the original to-be deleted node. 

Again, why not any other node? Because this subtree minimum will be either a leaf node or it will have only right branch. So we are reducing it to case 2. - that is node with only one branch or 0 branches. 

Next let us look at the code. I have written a recursive function here. You can also a find function to find tree minimum.

The advantage of recursive function is, you don't have to write separate functions to search a node, find the parent node etc.  Of course at the expense of simplicity.

NODEPTR delete_node(NODEPTR nd,int delval)
{
 if(nd==NULL)
          return nd;
 if(nd->val >delval)
           nd->left = delete_node(nd->left,delval);
        else if(nd->val < delval)
           nd->right = delete_node(nd->right,delval);
        else/*node found*/
        {
           if(nd->left==NULL && nd->right==NULL)/*leaf node*/
               {
                   free(nd);
     nd = NULL;
               }
           else if(nd->left==NULL)/*only right child*/
              {
                 NODEPTR temp = nd->right;
                 free(nd);
                 nd =temp;
              }
           else if(nd->right==NULL)/*only left child*/
              {
                 NODEPTR temp = nd->left;
                 free(nd);
                 nd = temp;
              }
           else/*both branches*/
            {
              NODEPTR min_node = find_rightst_min(nd->right);
              nd->val = min_node->val;/*copy value of minimumNode*/
              nd->right = delete_node(nd->right,min_node->val);/*delete minimumNode*/
            }         
      }
    return nd;  
}

/*go left till you reach null*/
NODEPTR find_rightst_min(NODEPTR nd)
{
     NODEPTR temp = nd;
     while(nd)
     {
        temp = nd;
        nd = nd->left;
      }
      return temp;
}

Now let us write the complete program where you keep deleting different nodes in a loop.


#include<stdio.h>
#include<stdlib.h>
struct node
{
   int val;
   struct node *left;
   struct node *right;
};
typedef struct node *NODEPTR;

NODEPTR create_node(int num)
{
     NODEPTR temp = (NODEPTR)malloc(sizeof(struct node));
     temp->val = num;
     temp->left = NULL;
     temp->right = NULL;
     return temp;
}

NODEPTR insert_node(NODEPTR nd,NODEPTR newnode)
{
    if(nd==NULL)
       return newnode;/* newnode becomes root of tree*/
    if(newnode->val > nd->val)
        nd->right = insert_node(nd->right,newnode);
    else if(newnode->val <  nd->val)
        nd->left = insert_node(nd->left,newnode); 
    return nd;   
}

void in_order(NODEPTR nd)
{
   if(nd!=NULL)
    {
        in_order(nd->left);
        printf("%d---",nd->val);
        in_order(nd->right);
     }
}

void pre_order(NODEPTR nd)
{
   if(nd!=NULL)
    {
        printf("%d---",nd->val);   
        pre_order(nd->left);
        pre_order(nd->right);
            
     }
}

void post_order(NODEPTR nd)
{
   if(nd!=NULL)
    {    
 post_order(nd->left);
        post_order(nd->right);
        printf("%d---",nd->val);
     }
}


/*go left till you reach null*/
NODEPTR find_rightst_min(NODEPTR nd)
{
     NODEPTR temp = nd;
     while(nd)
     {
        temp = nd;
        nd = nd->left;
      }
      return temp;
}
NODEPTR delete_node(NODEPTR nd,int delval)
{
 if(nd==NULL)
          return nd;
 if(nd->val >delval)
           nd->left = delete_node(nd->left,delval);
        else if(nd->val < delval)
           nd->right = delete_node(nd->right,delval);
        else/*node found*/
        {
           if(nd->left==NULL && nd->right==NULL)
               {
                   free(nd);
     nd = NULL;
               }
           else if(nd->left==NULL)/*leaf node*/
              {
                 NODEPTR temp = nd->right;
                 free(nd);
                 nd =temp;
              }
           else if(nd->right==NULL)/*leaf node*/
              {
                 NODEPTR temp = nd->left;
                 free(nd);
                 nd = temp;
              }
           else
            {
              NODEPTR min_node = find_rightst_min(nd->right);
              nd->val = min_node->val;
              nd->right = delete_node(nd->right,min_node->val);
            }
         
      }
return nd;
 
  
}
int main()
{
       NODEPTR root=NULL,delnode; 
       int n;
       do
       {
           NODEPTR newnode;
           printf("Enter value of node(-1 to exit):");
           scanf("%d",&n);
           if(n!=-1)
            {  
               newnode = create_node(n);
               root = insert_node(root,newnode);
             }
       } while (n!=-1);
       
       printf("\nInorder traversal\n");
       in_order(root);
      
       
       while(1){ 
   printf("Enter node to be deleted(-1 to stop)");
         scanf("%d",&n);
  if(n==-1)
   break;
           root = delete_node(root,n);
           printf("now tree is");
  in_order(root);
       }    
       return 0;
}
Labels:

Comments

Post a Comment

Popular posts from this blog

Linked list in C++

A linked list is a versatile data structure. In this structure, values are linked to one another with the help of addresses. I have written in an earlier post about how to create a linked list in C.  C++ has a library - standard template library which has list, stack, queue etc. data structures. But if you were to implement these data structures yourself in C++, how will you implement? If you just use new, delete, cout and cin, and then claim it is your c++ program, you are not conforming to OOPS concept. Remember you have to "keep it together". Keep all the functions and variables together - in a class. You have to have class called linked list in which there are methods - append, delete, display, insert, find, find_last. And there will also be a data - head. Defining node We need a structure for all these nodes. A struct can be used for this purpose, just like C. struct node { int val; struct node * next; }; Next we need to define our class. W
Post a Comment
Read more

Swap nodes of a linked list

Qn: Write a function to swap the adjacent nodes of a singly linked list.i.e. If the list has nodes as 1,2,3,4,5,6,7,8, after swapping, the list should be 2,1,4,3,6,5,8,7 Image from: https://tekmarathon.com Though the question looks simple enough, it is tricky because you don't just swap the pointers. You need to take care of links as well. So let us try to understand how to go about it. Take two adjacent nodes p1 and p2 Let prevnode be previous node of p1 Now link prevnode to p2 Link p2 to p1 Link p1 to next node of p2 So the code will be prevnode -> next = p2; p1 -> next = p2 -> next; p2 -> next = p1; But what about the start node or head? head node does not have previous node If we swap head with second node, modified head should be sent back to caller  To take care of swapping first and second nodes, we can write p1 = head; p2 = head -> next; p1 -> next = p2 -> next; p2 -> next = p1; head = p2;  Now we are read
Post a Comment
Read more

Introduction to AVL tree

AVL tree is a balanced binary search tree where the difference between heights of two sub trees is maximum 1. Why balanced tree A binary tree is good data structure because search operation here is of the order of O(logn). But this is true if the tree is balanced - which means the left and right subtrees are almost equal in height. If not balanced, search operation will take longer.  In worst case, if the tree has only one branch, then search is of the order O(n). Look at this example.  Here all nodes have only right children.  To search a value in this tree, we need upto 7 iterations, which is O(n). So this tree is very very inefficient. One way of making the tree efficient is to, balance the tree and make sure that height of two branches of each node are almost equal. Height of a node Height of a node is the distance between the node and its extreme child.  In the above a diagram, height of 37 is 3 and height of left child of 37 is 0 and right child of 37 is 2. Balancing a node A nod
Post a Comment
Read more