Go语言实现LRU算法

LRU 通常使用hash map + doubly linked list实现。在Golange中很简单，使用List保存数据,Map来做快速访问即可.

具体实现了下面几个函数:

func NewLRUCache(cap int)(*LRUCache)
func (lru *LRUCache)Set(k,v interface{})(error)
func (lru *LRUCache)Get(k interface{})(v interface{},ret bool,err error)
func (lru *LRUCache)Remove(k interface{})(bool)

演示:

package main

//LRU Cache
//author:Xiong Chuan Liang
//date:2015-2-3

import (
	"fmt"
	"github.com/xcltapestry/xclpkg/algorithm"
)

func main(){

	lru := algorithm.NewLRUCache(3)

	lru.Set(10,"value1")
	lru.Set(20,"value2")
	lru.Set(30,"value3")
	lru.Set(10,"value4")
	lru.Set(50,"value5")

	fmt.Println("LRU Size:",lru.Size())
	v,ret,_ := lru.Get(30)
	if ret  {
		fmt.Println("Get(30) : ",v)
	}

	if lru.Remove(30) {
		fmt.Println("Remove(30) : true ")
	}else{
		fmt.Println("Remove(30) : false ")
	}
	fmt.Println("LRU Size:",lru.Size())
}

运行结果:

LRU Size: 3
Get(30) :  value3
Remove(30) : true
LRU Size: 2

具体的 实现源码:

package algorithm

//LRU Cache
//author:Xiong Chuan Liang
//date:2015-2-3
//"github.com/xcltapestry/xclpkg/algorithm"  

import (
	"container/list"
	"errors"
)

type CacheNode struct {
	Key,Value interface{}
}

func (cnode *CacheNode)NewCacheNode(k,v interface{})*CacheNode{
	return &CacheNode{k,v}
}

type LRUCache struct {
	Capacity int
	dlist *list.List
	cacheMap map[interface{}]*list.Element
}

func NewLRUCache(cap int)(*LRUCache){
	return &LRUCache{
				Capacity:cap,
				dlist: list.New(),
				cacheMap: make(map[interface{}]*list.Element)}
}

func (lru *LRUCache)Size()(int){
	return lru.dlist.Len()
}

func (lru *LRUCache)Set(k,v interface{})(error){

	if lru.dlist == nil {
		return errors.New("LRUCache结构体未初始化.")
	}

	if pElement,ok := lru.cacheMap[k]; ok {
		lru.dlist.MoveToFront(pElement)
		pElement.Value.(*CacheNode).Value = v
		return nil
	}

	newElement := lru.dlist.PushFront( &CacheNode{k,v} )
	lru.cacheMap[k] = newElement

	if lru.dlist.Len() > lru.Capacity {
		//移掉最后一个
		lastElement := lru.dlist.Back()
		if lastElement == nil {
			return nil
		}
		cacheNode := lastElement.Value.(*CacheNode)
		delete(lru.cacheMap,cacheNode.Key)
		lru.dlist.Remove(lastElement)
	}
	return nil
}

func (lru *LRUCache)Get(k interface{})(v interface{},ret bool,err error){

	if lru.cacheMap == nil {
		return v,false,errors.New("LRUCache结构体未初始化.")
	}

	if pElement,ok := lru.cacheMap[k]; ok {
		lru.dlist.MoveToFront(pElement)
		return pElement.Value.(*CacheNode).Value,true,nil
	}
	return v,false,nil
}

func (lru *LRUCache)Remove(k interface{})(bool){

	if lru.cacheMap == nil {
		return false
	}

	if pElement,ok := lru.cacheMap[k]; ok {
		cacheNode := pElement.Value.(*CacheNode)
		delete(lru.cacheMap,cacheNode.Key)
		lru.dlist.Remove(pElement)
		return true
	}
	return false
}

附注:

1.key记录在map

2.对于set/get添加或命中的元素移到链表头

3.如总个数大于Cache容量(cap),则将最末的元素移除.

MAIL: [email protected]

BLOG:http://blog.csdn.net/xcl168