LCA
LCA of BST
- LCA 问题就是判断给定两个 node {p,q} 与 root 之间的相对位置。在 BST 里这种相对关系看 node.val 就可以。
- 时间复杂度 O(n)
- Worst case 如果 Tree 的形状是一条线往左或右的 full binary tree 的话。
- 因为是尾递归(递归后没有别的操作,例如说不需要回溯),显而易见的改法是用 while循环省去递归占用的系统栈空间;
public class Solution { public TreeNode lowestCommonAncestor(TreeNode root, TreeNode p, TreeNode q) { // Got to check if p and q is null as well; while(root != null){ if(root.val > p.val && root.val > q.val) root = root.left; else if(root.val < p.val && root.val < q.val) root = root.right; else return root; } return root; } }
LCA of BT
Given the root and two nodes in a Binary Tree. Find the lowest common ancestor(LCA) of the two nodes. The lowest common ancestor is the node with largest depth which is the ancestor of both nodes.
Analysis
- Binary Tree -> Divide & Conquer 所以考虑的时候,应该从如何将问题分解成子问题入手,而非直接去找解决问题的方法(通项)
- Divide 分解子问题,一般都有返回值
- 分析的时候可以画图,发现从root开始往下搜索,先搜索左子树,如果左边遇到任何==p或者q的值,则可以把这个值返回到当前的root,然后对另一侧(即右子树)进行搜索,如果右字数也遇到p或者q的值,则返回到root,这时候对于root的搜索,左右子树均有返回值,说明root就是他们的LCA
- 如果右边的返回值为null,说明在root节点的搜索只能找到一个点,需要往上一层搜索另外一侧

- 不管如何都需要搜索整个树的所有节点
- 这道题的tricky之处在于,理解返回值与LCA的定义
- if both p and q exist in Tree rooted at root, then return their LCA
- if neither p and q exist in Tree rooted at root, then return null
- if only one of p or q (NOT both of them), exists in Tree rooted at root, return it
- 如果右边的返回值为null,说明在root节点的搜索只能找到一个点,需要往上一层搜索另外一侧
- 分析的时候可以画图,发现从root开始往下搜索,先搜索左子树,如果左边遇到任何==p或者q的值,则可以把这个值返回到当前的root,然后对另一侧(即右子树)进行搜索,如果右字数也遇到p或者q的值,则返回到root,这时候对于root的搜索,左右子树均有返回值,说明root就是他们的LCA
- corner case
- 当A或B不在tree里
- A或B为root
- 这种解法的时间复杂度是 O(n),因为对于一个 node 来讲,它只被调用一次
- 极少会有 O(n^2) 的 binary tree 算法,因为那意味着每个节点相对于整个树重新计算,而不再是自己从属的路径或者高度。
Solution
public class Solution {
// 在root为根的二叉树中找A,B的LCA:
// 如果A和B都在root里,返回root
// 如果只有A在root里,就返回A
// 如果只有B在root里,就返回B
// 如果都没有,就返回null
public TreeNode lowestCommonAncestor(TreeNode root, TreeNode node1, TreeNode node2) {
if (root == null || root == node1 || root == node2) {
return root;
}
// Divide
TreeNode left = lowestCommonAncestor(root.left, node1, node2);
TreeNode right = lowestCommonAncestor(root.right, node1, node2);
// Conquer
if (left != null && right != null) {
return root;
}
if (left != null) {
return left;
}
if (right != null) {
return right;
}
return null;
}
}
Time Complexity: O(n)
Follow ups
if node has parent point
Analysis
分别获取A和B到root的paths,再根据path倒序比较,遇到第一个不同的则跳出,前一个就是LCA
public ParentTreeNode lowestCommonAncestorII(ParentTreeNode root, ParentTreeNode A, ParentTreeNode B) {
ArrayList<ParentTreeNode> pathA = getPath2Root(A);
ArrayList<ParentTreeNode> pathB = getPath2Root(B);
int indexA = pathA.size() - 1;
int indexB = pathB.size() - 1;
ParentTreeNode lowestAncestor = null;
while (indexA >= 0 && indexB >= 0) {
if (pathA.get(indexA) == pathB.get(indexB)) {
break;
}
lowestAncestor = pathA.get(indexA);
indexA--;
indexB--;
}
return lowestAncestor;
}
private ArrayList<ParentTreeNode> getPath2Root(ParentTreeNode node) {
ArrayList<ParentTreeNode> path = new ArrayList<>();
while (node != null) {
path.add(node);
node = node.parent;
}
return path;
}
if A or B may not be in the tree
Analysis
- 引入ResultType,存放A或者B的是否在树上进行判断
class ResultType {
public boolean a_exist, b_exist;
public TreeNode node;
ResultType(boolean a, boolean b, TreeNode n) {
a_exist = a;
b_exist = b;
node = n;
}
}
public class Solution {
/**
* @param root The root of the binary tree.
* @param A and B two nodes
* @return: Return the LCA of the two nodes.
*/
public TreeNode lowestCommonAncestor3(TreeNode root, TreeNode A, TreeNode B) {
// write your code here
ResultType rt = helper(root, A, B);
if (rt.a_exist && rt.b_exist)
return rt.node;
else
return null;
}
public ResultType helper(TreeNode root, TreeNode A, TreeNode B) {
if (root == null)
return new ResultType(false, false, null);
ResultType left_rt = helper(root.left, A, B);
ResultType right_rt = helper(root.right, A, B);
boolean a_exist = left_rt.a_exist || right_rt.a_exist || root == A;
boolean b_exist = left_rt.b_exist || right_rt.b_exist || root == B;
if (root == A || root == B)
return new ResultType(a_exist, b_exist, root);
if (left_rt.node != null && right_rt.node != null)
return new ResultType(a_exist, b_exist, root);
if (left_rt.node != null)
return new ResultType(a_exist, b_exist, left_rt.node);
if (right_rt.node != null)
return new ResultType(a_exist, b_exist, right_rt.node);
return new ResultType(a_exist, b_exist, null);
}
}