binary search tree impelementation and java

Question

I am trying to implement BST algorithm using Cormen\'s pseudo code yet having issue.

Here is my Code for Node:

public class Node {
    Node left;
            


        

Answer 1

Some immediate problems with your code: your treeSuccessor starts with

    if (x.right == null){
        return treeMinimum(x.right);
    }

which should be if (x.right != null), of course.

Your insert code has the lines

    Node tmp = getParent(t,z);
    tmp = y;

where you assign to tmp and immediately assign to it again. It doesn't seem to me that you need these lines at all, since you don't use tmp further on. At this moment, you have y being the node to whose child z gets inserted, so just delete these lines.

Again, in delete, you have the lines

    if (x != null){
        Node tmp = getParent(t,x);
        tmp = getParent(t,y);
    }

where you don't actually do anything, since tmp is not visible outside this snippet. And further on, in delete, you repeat the expression getParent(t,y), which can be an expensive operation, so you should compute it only once and assign it to some variable.

But in general, your code, though it seems correct (probably apart from delete, which I did not understand completely but which looks suspicious), does not much resemble typical binary tree code. You don't really need the getParent and treeSuccessor methods to implement search, insert, and delete. The basic structure that you have for search works for the others too, with the following modifications:

• with insert, when you get to a null link, instead of returning null, insert the element to that point
• with delete, when you find the element, if it has only one (or no) child, replace it with that child, and if it has two children, replace it with either the maximum of the left child tree or the minimum of the right child tree

Both of these require in addition that you keep track of the parent node while descending into the tree, but that's the only modification you need to make to search. In particular, there is never any need to go upwards in the tree (which treeSuccessor will do).