Why do we need multiple constructors for defining a binary tree in Java?

Question

Here is my code for defining a binary tree in Java

// Class definition for binary tree 
// Class for storing a single node of a binary tree of ints
// Cannot be set to be private because this should be accessed by the public class method

public class IntTreeNode {
    public int data;
    public IntTreeNode left;
    public IntTreeNode right;

    // constructs a leaf node with given data
    public IntTreeNode(int data) {
        this(data, null, null);
    }

    // constructs a branch node with given data, left subtree,
    // right subtree
    public IntTreeNode(int data, IntTreeNode left, 
                       IntTreeNode right) {
        this.data = data;
        this.left = left;
        this.right = right;
    }
}

Could I ask why we should have two constructors, public IntTreeNode(int data) and public IntTreeNode(int data, IntTreeNode left, IntTreeNode right) ? I ran the code without public IntTreeNode(int data) from a client class, it also works.

Answer 1

They do different things. One makes a new node with no children, one makes a new node with 2 children. While the second will work for both, it provides a more convenient and easier to read api.

Answer 2

Running the code simply with only public IntTreeNode(int data, IntTreeNode left, IntTreeNode right) would be perfectly legal code.

The 2 constructors here are for convenience. For a leaf node, it is not possible to have subtrees. Therefore, the one-parameter constructor provides a convenient way to add a leaf node, with the subtrees defaulted to null . On the other hand, any non-leaf node can have up to 2 subtrees, and so the second constructor exists to create such nodes.

Answer 3

Each constructor has its own role.

First constructor is used when creating a tree which only has a single root node.

Second constructor is used when creating a tree with left and right nodes. Those nodes can be other trees so you can use the second constructor to connect multiple trees.

Answer 4

Technically you are correct, only the second method is needed. But it's a lot clearer when written as two functions, one for leaves, the other for branch nodes. Otherwise the caller has to be told to call IntTreeNode as data,null,null for a leaf.