1.红黑树的概念
是一种二叉搜索树,在每个节点上增加一个存储位表示节点的颜色,Red或black,通过对任何一条从根到叶子的路径上各个结点着色方式的限制,确保没有一条路径会比其他路径长出俩倍,是接近平衡的。
2.红黑树的性质
- 每个结点不是红色就是黑色
- 根节点是黑色的
- 如果一个节点是红色的,则它的两个孩子结点必须是黑色的 ,任何路径没有连续的红色节点
- 对于每个结点,从该结点到其所有后代叶结点的简单路径上,每条路径上黑色节点数量相等
- 每个NIL叶子结点都是黑色的
为什么满足上面的性质,红黑树就能保证:其最长路径中节点个数不会超过最短路径节点个数的两倍?
通过该图结合性质3,4就能满足
3.红黑树的实现
3.1节点的定义
enum Colur
{
RED,
BLACK
};
template<class k,class v>
struct RBTreeNode
{
RBTreeNode<k, v>* _left;
RBTreeNode<k, v>* _right;
RBTreeNode<k, v>* _parent;
pair<k, v> _kv;
Colur _col;
BLTreeNode(const pair<k, v>& kv)
:_left(nullptr)
, _right(nullptr)
, _parent(nullptr)
,_kv(kv)
,_col(RED)
{ }
};
与AVL树相比,将平衡因子替换成颜色。
为什么默认将新插入节点颜色给成红色?
因为给黑色导致该路径黑色节点增加,红黑树性质每条路径上黑色节点数量相同,这样会导致所有路径节点发生变化。而给红色,当前路径违反了不能有两个连续红色节点性质,进行局部颜色调整和旋转即可。
3.2插入操作
- 按照二叉搜索的树规则插入新节点
- 检测新节点插入后,红黑树的性质是否造到破坏
约定cur为当前节点,p为父节点,g为祖父节点,u为叔叔节点
- 情况一:cur为红,p为红,g为黑,u存在且为红
这是一般规律下的抽象图,下面用一个特例图来帮助我们更好理解情况1的实现过程
- 情况二:cur为红,p为红,g为黑,u不存在/u存在且为黑
总结:
红黑树插入关键看uncle
1.uncle存在且为红,变色+继续往上更新
2.uncle不存在,uncle存在且为黑,旋转+变色
bool Insert(const pair<k, v>& kv)
{
if (_root == nullptr)
{
_root = new Node(kv);
_root->_col = BLACK;
return true;
}
//查找插入位置
Node* cur = _root;
Node* parent = nullptr;
while (cur)
{
if (cur->_kv.first < kv.first)
{
parent = cur;
cur = cur->_right;
}
else if (cur->_kv.first > kv.first)
{
parent = cur;
cur = cur->_left;
}
else
{//找到相同键值
return false;
}
}
//插入新节点
cur = new Node(kv);
cur->_col = RED;
if (parent->_kv.first > kv.first)
{
parent->_left = cur;
}
else
{
parent->_right = cur;
}
cur->_parent = parent;
while (parent&&parent->_col==RED)
{
Node* grandfather = parent->_parent;
if (parent == grandfather->_left)
{
Node* uncle = grandfather->_right;
//uncle存在且为红
if (uncle && uncle->_col == RED)
{
//变色
parent->_col = uncle->_col = BLACK;
grandfather->_col = RED;
//更新节点.继续向上处理
cur= grandfather;//注意赋值的顺序
parent = grandfather->_parent;
}
else//uncle不存在或为黑
{//判断是单旋还是双旋
if (cur == parent->_left)
{
// g
// p
//c
RotateR(grandfather);
parent->_col = BLACK;
grandfather->_col = RED;
}
else
{
// g
// p
// c
RotateL(parent);
RotateR(grandfather);
cur->_col = BLACK;
grandfather->_col = RED;
}
break;
}
}
else//parent == grandfather->_right
{
Node* uncle = grandfather->_left;
//uncle存在且为红
if (uncle && uncle->_col == RED)
{
parent->_col = uncle->_col = BLACK;
grandfather->_col = RED;
//更新节点.继续向上处理
cur= grandfather;
parent = grandfather->_parent;
}
else//uncle不存在或为黑
{//判断是单旋还是双旋
if (cur == parent->_right)
{
//g
// p
// c
RotateL(grandfather);
parent->_col = BLACK;
grandfather->_col = RED;
}
else
{
//g
// p
//c
RotateR(parent);
RotateL(grandfather);
cur->_col = BLACK;
grandfather->_col = RED;
}
break;
}
}
}
_root->_col = BLACK;
return true;
}
代码思路:
1.查找插入位置,创建节点并插入,进行平衡调整。
2.平衡调整中循环成立条件为:父节点存在且颜色为红。因为新插入节点为红,父节点也为红,意味着存在两个连续的红色节点,违反红黑树性质,此时需要通过变色和旋转来修复树的性质。
3.在循环中确定祖父节点和叔叔节点。通过判断叔叔节点来进行操作,如上文总结。
4.当父节点不存在或为黑色时调整结束。
3.3验证
- 检测其是否满足二叉搜索树(中序遍历是否为有序序列)
将红黑树按中序遍历的方式存入vector中,通过前后比较元素看是否满足升序
bool IsBST() {
vector<int> result;
_Inorder(_root,result);
for (int i = 1; i < result.size(); i++) {
if (result[i] <= result[i - 1]) {
return false;
}
}
return true;
}
void _Inorder(Node* root, vector<k>& result) {
if (root == nullptr) {
return;
}
_Inorder(root->_left, result);
result.push_back(root->_kv.first);
_Inorder(root->_right, result);
}
2.检测是否满足红黑树性质
- 检查颜色
bool CheckColur(Node* root, int blacknum, int benchmark)
{//走到null后判断黑色节点数量与基准值是否相同
if (root == nullptr)
{//判断是否违反每条路径中黑色节点数量必须相同
if (blacknum != benchmark)
{
return false;
}
return true;
}
//判断树中黑节点数量
if (root->_col==BLACK)
{
blacknum++;
}
//判断树中是否有连续红节点情况
if (root->_col == RED && root->_parent && root->_parent->_col == RED)
{
cout << root->_kv.first << "出现连续红色节点" << endl;
return false;
}
return CheckColur(root->_left, blacknum, benchmark)
&& CheckColur(root->_right, blacknum, benchmark);
}
注意:
这里第二个参数blacknum不传引用刚刚好,这样可以正确的递归计算每一条路径黑色节点数量,返回时由于不是传引用,blacknum作为临时变量随着栈帧的销毁而销毁,不会影响到另一条路径黑色节点数量的计算
//封装接口,提供外界使用
bool IsBalance()
{
return IsBalance(_root);
}
bool IsBalance(Node* root)
{
if (root == nullptr)
{
return true;
}
if (root->_col != BLACK)
{
return false;
}
//计算基准值,随便计算一条路径黑节点的数量
int benchmark = 0;
Node* cur = root;
while (cur)
{
if(cur->_col==BLACK)
benchmark++;
//选择一条路径来计算黑节点数量
cur = cur->_left;
}
return CheckColur(root, 0, benchmark);
}
3.4红黑树与AVL树的比较
都是高效的平衡二叉树,增删改查的时间复杂度都是O( l o g 2 N log_2 N log2N),红黑树不追求绝对平衡,其只需保证最长路径不超过最短路径的2倍,相对而言,降低了插入和旋转的次数,所以在经常进行增删的结构中性能比AVL树更优,而且红黑树实现比较简单,所以实际运用中红黑树更多。
4.整体代码
- RBTree.h
#pragma once
#include<iostream>
#include<vector>
using namespace std;
enum Colur
{
RED,
BLACK
};
template<class k,class v>
struct RBTreeNode
{
RBTreeNode<k, v>* _left;
RBTreeNode<k, v>* _right;
RBTreeNode<k, v>* _parent;
pair<k, v> _kv;
Colur _col;
RBTreeNode(const pair<k, v>& kv)
:_left(nullptr)
, _right(nullptr)
, _parent(nullptr)
,_kv(kv)
,_col(RED)
{ }
};
template<class k,class v>
struct RBTree
{
typedef RBTreeNode<k, v> Node;
public:
bool Insert(const pair<k, v>& kv)
{
if (_root == nullptr)
{
_root = new Node(kv);
_root->_col = BLACK;
return true;
}
//查找插入位置
Node* cur = _root;
Node* parent = nullptr;
while (cur)
{
if (cur->_kv.first < kv.first)
{
parent = cur;
cur = cur->_right;
}
else if (cur->_kv.first > kv.first)
{
parent = cur;
cur = cur->_left;
}
else
{//找到相同键值
return false;
}
}
//插入新节点
cur = new Node(kv);
cur->_col = RED;
if (parent->_kv.first > kv.first)
{
parent->_left = cur;
}
else
{
parent->_right = cur;
}
cur->_parent = parent;
while (parent&&parent->_col==RED)
{
Node* grandfather = parent->_parent;
if (parent == grandfather->_left)
{
Node* uncle = grandfather->_right;
//uncle存在且为红
if (uncle && uncle->_col == RED)
{
//变色
parent->_col = uncle->_col = BLACK;
grandfather->_col = RED;
//更新节点.继续向上处理
cur= grandfather;//注意赋值的顺序
parent = grandfather->_parent;
}
else//uncle不存在或为黑
{//判断是单旋还是双旋
if (cur == parent->_left)
{
// g
// p
//c
RotateR(grandfather);
parent->_col = BLACK;
grandfather->_col = RED;
}
else
{
// g
// p
// c
RotateL(parent);
RotateR(grandfather);
cur->_col = BLACK;
grandfather->_col = RED;
}
break;
}
}
else//parent == grandfather->_right
{
Node* uncle = grandfather->_left;
//uncle存在且为红
if (uncle && uncle->_col == RED)
{
parent->_col = uncle->_col = BLACK;
grandfather->_col = RED;
//更新节点.继续向上处理
cur= grandfather;
parent = grandfather->_parent;
}
else//uncle不存在或为黑
{//判断是单旋还是双旋
if (cur == parent->_right)
{
//g
// p
// c
RotateL(grandfather);
parent->_col = BLACK;
grandfather->_col = RED;
}
else
{
//g
// p
//c
RotateR(parent);
RotateL(grandfather);
cur->_col = BLACK;
grandfather->_col = RED;
}
break;
}
}
}
_root->_col = BLACK;
return true;
}
void RotateL(Node* parent)
{
_rotateCount++;
Node* cur = parent->_right;
Node* curleft = cur->_left;
Node* ppnode = parent->_parent;
//第一次改变链接
parent->_right = curleft;
if (curleft)
{
curleft->_parent = parent;
}
//第二次改变链接
cur->_left = parent;
parent->_parent = cur;
//判断根节点的链接情况
//为根节点调整平衡因子情况
if (parent == _root)
{
_root = cur;
cur->_parent = nullptr;
}
//树中的部分调整情况
else
{
if (ppnode->_left == parent)
{
ppnode->_left = cur;
}
else
{
ppnode->_right = cur;
}
cur->_parent = ppnode;
}
}
void RotateR(Node* parent)
{
_rotateCount++;
Node* cur = parent->_left;
Node* curright = cur->_right;
Node* ppnode = parent->_parent;
//第一次链接
parent->_left = curright;
if (curright)
{
curright->_parent = parent;
}
//第二次链接
cur->_right = parent;
parent->_parent = cur;
//调整根节点链接关系
if (parent == _root)
{
_root = cur;
cur->_parent = nullptr;
}
else
{
if (ppnode->_left == parent)
{
ppnode->_left = cur;
}
else
{
ppnode->_right = cur;
}
cur->_parent = ppnode;
}
}
bool CheckColur(Node* root, int blacknum, int benchmark)
{
if (root == nullptr)
{
if (blacknum != benchmark)
{
return false;
}
return true;
}
//判断树中黑节点数量
if (root->_col==BLACK)
{
blacknum++;
}
//判断树中是否有连续红节点情况
if (root->_col == RED && root->_parent && root->_parent->_col == RED)
{
cout << root->_kv.first << "出现连续红色节点" << endl;
return false;
}
return CheckColur(root->_left, blacknum, benchmark)
&& CheckColur(root->_right, blacknum, benchmark);
}
bool IsBalance()
{
return IsBalance(_root);
}
bool IsBalance(Node* root)
{
if (root == nullptr)
{
return true;
}
if (root->_col != BLACK)
{
return false;
}
//计算基准值
int benchmark = 0;
Node* cur = root;
while (cur)
{
if(cur->_col==BLACK)
benchmark++;
//选择一条路径来计算黑节点数量
cur = cur->_left;
}
return CheckColur(root, 0, benchmark);
}
int Height()
{
return Height(_root);
}
int Height(Node* root)
{
if (root == nullptr)
{
return 0;
}
int leftheight = Height(root->_left);
int rightheight = Height(root->_right);
return leftheight > rightheight ?
leftheight + 1 : rightheight + 1;
}
bool IsBST() {
vector<int> result;
_Inorder(_root,result);
for (int i = 1; i < result.size(); i++) {
if (result[i] <= result[i - 1]) {
return false;
}
}
return true;
}
void _Inorder(Node* root, vector<k>& result) {
if (root == nullptr) {
return;
}
_Inorder(root->_left, result);
result.push_back(root->_kv.first);
_Inorder(root->_right, result);
}
public:
int _rotateCount = 0;
private:
Node* _root = nullptr;
};
- 测试代码
#include"AVLTree.h"
#include"RBTree.h"
#include<vector>
//int main()
//{
// //int a[] = { 16, 3, 7, 11, 9, 26, 18, 14, 15 };
// int a[] = { 4, 2, 6, 1, 3, 5, 15, 7, 16, 14 };
// RBTree<int, int> t;
// for (auto e : a)
// {
// t.Insert(make_pair(e, e));
// cout << "Insert:" << e << "->" << t.IsBalance() << endl;
// }
//
// return 0;
//}
int main()
{
const int N = 1000000;
vector<int>v;
srand(time(0));
v.reserve(N);
for (size_t i = 0; i < N; i++)
{
v.push_back(i);
}
AVLTree<int, int>avlt;
for (auto e : v)
{
avlt.Insert(make_pair(e, e));
}
cout << avlt.IsBalance() << endl;
cout << avlt.Height() << endl;
cout << avlt._rotateCount << endl;
cout << avlt.IsBST() << endl;
RBTree<int, int>rbt;
for (auto e : v)
{
rbt.Insert(make_pair(e, e));
}
cout << rbt.IsBalance() << endl;
cout << rbt.Height() << endl;
cout << rbt._rotateCount << endl;
cout << rbt.IsBST() << endl;
return 0;
}