Leetcode:LRUCache四个版本实现

时间:2024-04-08 10:07:07

题目

Design and implement a data structure for Least Recently Used (LRU) cache. It should support the following operations: get and set.

get(key) - Get the value (will always be positive) of the key if the key exists in the cache, otherwise return -1.
set(key, value) - Set or insert the value if the key is not already present. When the cache reached its capacity, it should invalidate the least recently used item before inserting a new item.

链接:https://oj.leetcode.com/problems/lru-cache/

分析

1:维护最近最少(LRU)使用的cache

  1)使用count计数,每次操作cache时(get、set),该cache的count置0,其余cache的count加1,count最大的为最近最少使用的cache

2)使用链表,每次操作cache时(get、set),将该cache移动至链表头,链表尾的cache为最近最少使用的cache

2:快速查找cache

  1)使用stl::unordered_map存储cache地址(内部hashTable)

版本一

使用std::list维护LRU,链表中存储cache实际空间。

Runtime: 248ms。

 #include <list>
#include <unordered_map> struct KeyValue {
KeyValue(int pKey, int pValue):key(pKey), value(pValue) {};
int key;
int value;
}; class LRUCache{
public:
LRUCache(int capacity):_capacity(capacity) {}; int get(int key);
void set(int key, int value); private:
std::unordered_map<int, std::list<KeyValue>::iterator> keyToNodeItr;
std::list<KeyValue> lru; int _capacity;
}; void LRUCache::set(int key, int value) {
auto itr = keyToNodeItr.find(key);
if (itr != keyToNodeItr.end()) { // set value
itr->second->value = value;
KeyValue tmp(itr->second->key, itr->second->value);
keyToNodeItr.erase(itr->second->key);
lru.erase(itr->second);
lru.push_front(tmp);
} else { // insert value
if (lru.size() != _capacity) {
lru.push_front(KeyValue(key, value));
} else {
// pop back lru
if (lru.size() != ) {
keyToNodeItr.erase((lru.rbegin())->key);
lru.pop_back();
}
lru.push_front(KeyValue(key, value));
}
} keyToNodeItr.insert(std::pair<int, std::list<KeyValue>::iterator>(key, lru.begin()));
} int LRUCache::get(int key) {
auto itr = keyToNodeItr.find(key);
if (itr != keyToNodeItr.end()) {
int value = itr->second->value; KeyValue tmp(itr->second->key, itr->second->value);
keyToNodeItr.erase(itr->second->key);
lru.erase(itr->second);
lru.push_front(tmp);
keyToNodeItr.insert(std::pair<int, std::list<KeyValue>::iterator>(key, lru.begin())); return value;
}
return -;
}

因为链表中存储的为cache的实际空间,因此当需要改变cache的位置时,链表及map都需要改变,开销较大。

版本二

使用std::list维护LRU,链表中存储cache的指针。

Runtime:186ms。

 #include <list>
#include <unordered_map> struct KeyValue {
KeyValue(int pKey, int pValue):key(pKey), value(pValue) {};
int key;
int value;
}; class LRUCache{
public:
LRUCache(int capacity):_capacity(capacity) {}; int get(int key);
void set(int key, int value); ~LRUCache(); private:
std::unordered_map<int, std::list<KeyValue*>::iterator> keyToNodeItr;
std::list<KeyValue*> lru; int _capacity;
}; LRUCache::~LRUCache() {
for (auto itr = lru.begin(); itr != lru.end(); ++itr) {
delete *itr;
}
} void LRUCache::set(int key, int value) {
auto itr = keyToNodeItr.find(key);
if (itr != keyToNodeItr.end()) { // set value
KeyValue* tmp = *(itr->second);
tmp->value = value;
lru.erase(itr->second);
lru.push_front(tmp);
itr->second = lru.begin(); // avoid invalid iterator
} else { // insert value
if (lru.size() == _capacity) { // pop back lru
KeyValue* tmp = *(lru.rbegin());
keyToNodeItr.erase(tmp->key);
delete tmp;
lru.pop_back();
}
lru.push_front(new KeyValue(key, value));
keyToNodeItr.insert(std::pair<int, std::list<KeyValue*>::iterator>(key, lru.begin()));
}
} int LRUCache::get(int key) {
auto itr = keyToNodeItr.find(key);
if (itr != keyToNodeItr.end()) {
KeyValue* kvPtr = *(itr->second);
lru.erase(itr->second);
lru.push_front(kvPtr);
itr->second = lru.begin();
return kvPtr->value;
}
return -;
}

需要注意的问题是map中存储的为list的迭代器,因此map中仍需要重新设置key到迭代器的映射,避免迭代器失效。

版本三

似乎std::list太笨重了,so实现轻量级双链表代替std::list。

Runtime: 77ms。

 struct BiListNode {
BiListNode() {};
BiListNode(int key, int value):key(key), value(value) {};
int key;
int value;
BiListNode* pre;
BiListNode* next;
}; class BiList {
public:
BiList():_count() {
_head = new BiListNode();
_head->pre = _head;
_head->next = _head;
} void push_front(BiListNode* pNode); void move_front(BiListNode* pNode); BiListNode* begin() {
return _head->next;
} BiListNode* rbegin() {
return _head->pre;
} void pop_back(); int size() { return _count; } ~BiList(); private:
BiListNode* _head;
int _count;
}; void BiList::push_front(BiListNode* pNode) {
pNode->next = _head->next;
pNode->pre = _head;
_head->next->pre = pNode;
_head->next = pNode;
if (_head->pre == _head) {
_head->pre = pNode;
}
++_count;
} void BiList::move_front(BiListNode* pNode) {
if (pNode == _head->next) {
return;
}
pNode->pre->next = pNode->next;
pNode->next->pre = pNode->pre; pNode->next = _head->next;
pNode->pre = _head; _head->next->pre = pNode; _head->next = pNode;
} void BiList::pop_back() {
BiListNode* tailPtr = _head->pre;
tailPtr->pre->next = _head;
_head->pre = tailPtr->pre;
delete tailPtr;
--_count;
} BiList::~BiList() {
for (BiListNode* itr = _head->next; itr != _head; itr = itr->next) {
delete itr;
}
delete _head;
} class LRUCache {
public:
LRUCache(int capacity):_capacity(capacity) {}; int get(int key); void set(int key, int value); private:
int _capacity; BiList biList;
std::unordered_map<int, BiListNode*> keyToNodePtr;
}; int LRUCache::get(int key) {
auto itr = keyToNodePtr.find(key);
if (itr != keyToNodePtr.end()) {
biList.move_front(itr->second);
return itr->second->value;
}
return -;
} void LRUCache::set(int key, int value) {
auto itr = keyToNodePtr.find(key);
if (itr != keyToNodePtr.end()) { // set value
itr->second->value = value;
biList.move_front(itr->second);
} else { // insert
if (biList.size() == _capacity) {
keyToNodePtr.erase((biList.rbegin())->key);
biList.pop_back();
}
biList.push_front(new BiListNode(key, value));
keyToNodePtr.insert(std::pair<int, BiListNode*>(key, biList.begin()));
}
}

自己实现的双链表仅有80行代码,代码运行效率大大提高。

版本四

双链表都自己实现了,就死磕到底,再自己实现个开链哈希表吧。

Runtime:66ms。

 struct BiListNode {
BiListNode() {};
BiListNode(int key, int value):key(key), value(value) {};
int key;
int value;
BiListNode* pre;
BiListNode* next;
}; class BiList {
public:
BiList():_count() {
_head = new BiListNode();
_head->pre = _head;
_head->next = _head;
} void push_front(BiListNode* pNode); void move_front(BiListNode* pNode); BiListNode* begin() {
return _head->next;
} BiListNode* rbegin() {
return _head->pre;
} void pop_back(); int size() { return _count; } ~BiList(); private:
BiListNode* _head;
int _count;
}; void BiList::push_front(BiListNode* pNode) {
pNode->next = _head->next;
pNode->pre = _head;
_head->next->pre = pNode;
_head->next = pNode;
if (_head->pre == _head) {
_head->pre = pNode;
}
++_count;
} void BiList::move_front(BiListNode* pNode) {
if (pNode == _head->next) {
return;
}
pNode->pre->next = pNode->next;
pNode->next->pre = pNode->pre; pNode->next = _head->next;
pNode->pre = _head; _head->next->pre = pNode; _head->next = pNode;
} void BiList::pop_back() {
BiListNode* tailPtr = _head->pre;
tailPtr->pre->next = _head;
_head->pre = tailPtr->pre;
delete tailPtr;
--_count;
} BiList::~BiList() {
for (BiListNode* itr = _head->next; itr != _head; itr = itr->next) {
delete itr;
}
delete _head;
} struct hashNode {
hashNode(int key, BiListNode* ptr):key(key), ptr(ptr), next(NULL) {};
int key;
BiListNode* ptr;
hashNode* next;
}; class HashTable {
public:
HashTable(int capacity); hashNode* find(int key); void insert(int key, BiListNode* ptr); void erase(int key); ~HashTable(); private:
int _capacity;
hashNode** hashArray;
}; HashTable::HashTable(int capacity):_capacity(capacity) {
hashArray = new hashNode*[capacity];
for (int i = ; i < _capacity; ++i) {
hashArray[i] = NULL;
}
} hashNode* HashTable::find(int key) {
for (hashNode* itr = hashArray[key % _capacity]; itr != NULL;
itr = itr->next) {
if (itr->key == key) {
return itr;
}
}
return NULL;
} void HashTable::insert(int key, BiListNode* ptr) {
hashNode* tmp = new hashNode(key, ptr); int relativeKey = key % _capacity; if (hashArray[relativeKey] == NULL) {
hashArray[relativeKey] = tmp;
return;
} tmp->next = hashArray[relativeKey];
hashArray[relativeKey] = tmp;
} void HashTable::erase(int key) {
for (hashNode* pre = hashArray[key % _capacity], *itr = pre;
itr != NULL; pre = itr, itr = itr->next) {
if (itr->key == key) {
if (itr != pre)
pre->next = itr->next;
else // head
hashArray[key % _capacity] = itr->next; delete itr;
}
}
} HashTable::~HashTable() {
for (int i = ; i < _capacity; ++i) {
for (hashNode* itr = hashArray[i]; itr != NULL;) {
hashNode* tmp = itr;
itr = itr->next;
delete tmp;
}
}
delete [] hashArray;
} class LRUCache {
public:
LRUCache(int capacity):_capacity(capacity) {
hashTable = new HashTable();
}; int get(int key); void set(int key, int value); ~LRUCache() { delete hashTable; } private:
int _capacity;
BiList bilist;
HashTable* hashTable;
}; int LRUCache::get(int key) {
hashNode* tmp = hashTable->find(key);
if (tmp != NULL) {
bilist.move_front(tmp->ptr);
return tmp->ptr->value;
}
return -;
} void LRUCache::set(int key, int value) {
hashNode* tmp = hashTable->find(key);
if (tmp != NULL) { // set
bilist.move_front(tmp->ptr);
tmp->ptr->value = value;
return;
} // insert
if (bilist.size() == _capacity) {
hashTable->erase((bilist.rbegin())->key);
bilist.pop_back();
} bilist.push_front(new BiListNode(key, value));
hashTable->insert(key, bilist.begin());
}

开链哈希表72行代码