LeetCode 94 Binary Tree Inorder Traversal

Given a binary tree, return the inorder traversal of its nodes' values.

For example:
Given binary tree [1,null,2,3],

   1
    \
     2
    /
   3

return [1,3,2].

Note: Recursive solution is trivial, could you do it iteratively?

通常，实现二叉树的前序（preorder）、中序（inorder）、后序（postorder）遍历有两个常用的方法：一是递归(recursive)，二是使用栈实现的迭代版本(stack+iterative)。这两种方法都是O(n)的空间复杂度（递归本身占用stack空间或者用户自定义的stack）

Recursive: Runtime: 1 ms  beats 60.97% of javasubmissions.

	public List<Integer> inorderTraversal(TreeNode root) {
		List<Integer> res = new ArrayList<Integer>();
		if (root == null) return res;
		res.addAll(inorderTraversal(root.left));
		res.add(root.val);
		res.addAll(inorderTraversal(root.right));
		return res;
	}

Iterative with Stack. Runtime: 2 ms  beats 4.67% of javasubmissions.

	public List<Integer> inorderTraversal2(TreeNode root) {
		List<Integer> results = new ArrayList<>();
		ArrayDeque<TreeNode> stack = new ArrayDeque<>();
		TreeNode crt = root;
		while (!stack.isEmpty() || crt != null) {
			if (crt != null) {
				stack.push(crt);
				crt = crt.left;
			} else {
				crt = stack.pop();
				results.add(crt.val);
				crt = crt.right;
			}
		}
		return results;
	}

Morris Traversal（threaded binary tree）方法遍历二叉树（非递归，不用栈，O(1)空间）中序遍历，时间复杂度 O(n).

Recursive:  Runtime:  1 ms  beats 60.97% of javasubmissions.

	public List<Integer> inorderTraversal3(TreeNode root) {
		List<Integer> results = new ArrayList<>();
		TreeNode crt = root;
		while (crt != null) {
			if (crt.left == null) {
				results.add(crt.val);
				crt = crt.right;
			} else {
				TreeNode temp = crt.left;/* 查找前驱 */
				while (temp.right != null && temp.right != crt)
					temp = temp.right;

				if (temp.right == null) {/* 还没线索化，则建立线索 */
					temp.right = crt;
					crt = crt.left;
				} else {/* 已经线索化，则访问节点，并删除线索 */
					temp.right = null;
					results.add(crt.val);
					crt = crt.right;
				}
			}
		}
		return results;
	}

参考http://www.cnblogs.com/AnnieKim/archive/2013/06/15/MorrisTraversal.html