• LRU缓存介绍与实现 (Java)


    LRU缓存:

    LRU缓存利用了这样的一种思想。LRU是Least Recently Used 的缩写,翻译过来就是“最近最少使用”,也就是说,LRU缓存把最近最少使用的数据移除,让给最新读取的数据。而往往最常读取的,也是读取次数最多的,所以,利用LRU缓存,我们能够提高系统的performance.

    实现:

    要实现LRU缓存,我们首先要用到一个类 LinkedHashMap。 用这个类有两大好处:一是它本身已经实现了按照访问顺序的存储,也就是说,最近读取的会放在最前面,最最不常读取的会放在最后(当然,它也可以实现按照插入顺序存储)。第二,LinkedHashMap本身有一个方法用于判断是否需要移除最不常读取的数,但是,原始方法默认不需要移除(这是,LinkedHashMap相当于一个linkedlist),所以,我们需要override这样一个方法,使得当缓存里存放的数据个数超过规定个数后,就把最不常用的移除掉。LinkedHashMap的API写得很清楚,推荐大家可以先读一下。

     1 import java.util.LinkedHashMap;  
     2 import java.util.Collection;  
     3 import java.util.Map;  
     4 import java.util.ArrayList;  
     5   
     6 /** 
     7 * An LRU cache, based on <code>LinkedHashMap</code>. 
     8 * 
     9 * <p> 
    10 * This cache has a fixed maximum number of elements (<code>cacheSize</code>). 
    11 * If the cache is full and another entry is added, the LRU (least recently used) entry is dropped. 
    12 * 
    13 * <p> 
    14 * This class is thread-safe. All methods of this class are synchronized. 
    15 * 
    16 * <p> 
    17 * Author: Christian d'Heureuse, Inventec Informatik AG, Zurich, Switzerland<br> 
    18 * Multi-licensed: EPL / LGPL / GPL / AL / BSD. 
    19 */  
    20 public class LRUCache<K,V> {  
    21   
    22 private static final float   hashTableLoadFactor = 0.75f;  
    23   
    24 private LinkedHashMap<K,V>   map;  
    25 private int                  cacheSize;  
    26   
    27 /** 
    28 * Creates a new LRU cache. 
    29 * @param cacheSize the maximum number of entries that will be kept in this cache. 
    30 */  
    31 public LRUCache (int cacheSize) {  
    32    this.cacheSize = cacheSize;  
    33    int hashTableCapacity = (int)Math.ceil(cacheSize / hashTableLoadFactor) + 1;  
    34    map = new LinkedHashMap<K,V>(hashTableCapacity, hashTableLoadFactor, true) {  
    35       // (an anonymous inner class)  
    36       private static final long serialVersionUID = 1;  
    37       @Override protected boolean removeEldestEntry (Map.Entry<K,V> eldest) {  
    38          return size() > LRUCache.this.cacheSize; }}; }  
    39   
    40 /** 
    41 * Retrieves an entry from the cache.<br> 
    42 * The retrieved entry becomes the MRU (most recently used) entry. 
    43 * @param key the key whose associated value is to be returned. 
    44 * @return    the value associated to this key, or null if no value with this key exists in the cache. 
    45 */  
    46 public synchronized V get (K key) {  
    47    return map.get(key); }  
    48   
    49 /** 
    50 * Adds an entry to this cache. 
    51 * The new entry becomes the MRU (most recently used) entry. 
    52 * If an entry with the specified key already exists in the cache, it is replaced by the new entry. 
    53 * If the cache is full, the LRU (least recently used) entry is removed from the cache. 
    54 * @param key    the key with which the specified value is to be associated. 
    55 * @param value  a value to be associated with the specified key. 
    56 */  
    57 public synchronized void put (K key, V value) {  
    58    map.put (key, value); }  
    59   
    60 /** 
    61 * Clears the cache. 
    62 */  
    63 public synchronized void clear() {  
    64    map.clear(); }  
    65   
    66 /** 
    67 * Returns the number of used entries in the cache. 
    68 * @return the number of entries currently in the cache. 
    69 */  
    70 public synchronized int usedEntries() {  
    71    return map.size(); }  
    72   
    73 /** 
    74 * Returns a <code>Collection</code> that contains a copy of all cache entries. 
    75 * @return a <code>Collection</code> with a copy of the cache content. 
    76 */  
    77 public synchronized Collection<Map.Entry<K,V>> getAll() {  
    78    return new ArrayList<Map.Entry<K,V>>(map.entrySet()); }  
    79   
    80 } // end class LRUCache  
    81 ------------------------------------------------------------------------------------------  
    82 // Test routine for the LRUCache class.  
    83 public static void main (String[] args) {  
    84    LRUCache<String,String> c = new LRUCache<String, String>(3);  
    85    c.put ("1", "one");                           // 1  
    86    c.put ("2", "two");                           // 2 1  
    87    c.put ("3", "three");                         // 3 2 1  
    88    c.put ("4", "four");                          // 4 3 2  
    89    if (c.get("2") == null) throw new Error();    // 2 4 3  
    90    c.put ("5", "five");                          // 5 2 4  
    91    c.put ("4", "second four");                   // 4 5 2  
    92    // Verify cache content.  
    93    if (c.usedEntries() != 3)              throw new Error();  
    94    if (!c.get("4").equals("second four")) throw new Error();  
    95    if (!c.get("5").equals("five"))        throw new Error();  
    96    if (!c.get("2").equals("two"))         throw new Error();  
    97    // List cache content.  
    98    for (Map.Entry<String, String> e : c.getAll())  
    99       System.out.println (e.getKey() + " : " + e.getValue()); }  

    代码出自:http://www.source-code.biz/snippets/java/6.htm


    在博客 http://gogole.iteye.com/blog/692103 里,作者使用的是双链表 + hashtable 的方式实现的。如果在面试题里考到如何实现LRU,考官一般会要求使用双链表 + hashtable 的方式。 所以,我把原文的部分内容摘抄如下:

    双链表 + hashtable实现原理:

    将Cache的所有位置都用双连表连接起来,当一个位置被命中之后,就将通过调整链表的指向,将该位置调整到链表头的位置,新加入的Cache直接加到链表头中。这样,在多次进行Cache操作后,最近被命中的,就会被向链表头方向移动,而没有命中的,而想链表后面移动,链表尾则表示最近最少使用的Cache。当需要替换内容时候,链表的最后位置就是最少被命中的位置,我们只需要淘汰链表最后的部分即可。

      1 public class LRUCache {  
      2       
      3     private int cacheSize;  
      4     private Hashtable<Object, Entry> nodes;//缓存容器  
      5     private int currentSize;  
      6     private Entry first;//链表头  
      7     private Entry last;//链表尾  
      8       
      9     public LRUCache(int i) {  
     10         currentSize = 0;  
     11         cacheSize = i;  
     12         nodes = new Hashtable<Object, Entry>(i);//缓存容器  
     13     }  
     14       
     15     /** 
     16      * 获取缓存中对象,并把它放在最前面 
     17      */  
     18     public Entry get(Object key) {  
     19         Entry node = nodes.get(key);  
     20         if (node != null) {  
     21             moveToHead(node);  
     22             return node;  
     23         } else {  
     24             return null;  
     25         }  
     26     }  
     27       
     28     /** 
     29      * 添加 entry到hashtable, 并把entry  
     30      */  
     31     public void put(Object key, Object value) {  
     32         //先查看hashtable是否存在该entry, 如果存在,则只更新其value  
     33         Entry node = nodes.get(key);  
     34           
     35         if (node == null) {  
     36             //缓存容器是否已经超过大小.  
     37             if (currentSize >= cacheSize) {  
     38                 nodes.remove(last.key);  
     39                 removeLast();  
     40             } else {  
     41                 currentSize++;  
     42             }             
     43             node = new Entry();  
     44         }  
     45         node.value = value;  
     46         //将最新使用的节点放到链表头,表示最新使用的.  
     47         moveToHead(node);  
     48         nodes.put(key, node);  
     49     }  
     50   
     51     /** 
     52      * 将entry删除, 注意:删除操作只有在cache满了才会被执行 
     53      */  
     54     public void remove(Object key) {  
     55         Entry node = nodes.get(key);  
     56         //在链表中删除  
     57         if (node != null) {  
     58             if (node.prev != null) {  
     59                 node.prev.next = node.next;  
     60             }  
     61             if (node.next != null) {  
     62                 node.next.prev = node.prev;  
     63             }  
     64             if (last == node)  
     65                 last = node.prev;  
     66             if (first == node)  
     67                 first = node.next;  
     68         }  
     69         //在hashtable中删除  
     70         nodes.remove(key);  
     71     }  
     72   
     73     /** 
     74      * 删除链表尾部节点,即使用最后 使用的entry 
     75      */  
     76     private void removeLast() {  
     77         //链表尾不为空,则将链表尾指向null. 删除连表尾(删除最少使用的缓存对象)  
     78         if (last != null) {  
     79             if (last.prev != null)  
     80                 last.prev.next = null;  
     81             else  
     82                 first = null;  
     83             last = last.prev;  
     84         }  
     85     }  
     86       
     87     /** 
     88      * 移动到链表头,表示这个节点是最新使用过的 
     89      */  
     90     private void moveToHead(Entry node) {  
     91         if (node == first)  
     92             return;  
     93         if (node.prev != null)  
     94             node.prev.next = node.next;  
     95         if (node.next != null)  
     96             node.next.prev = node.prev;  
     97         if (last == node)  
     98             last = node.prev;  
     99         if (first != null) {  
    100             node.next = first;  
    101             first.prev = node;  
    102         }  
    103         first = node;  
    104         node.prev = null;  
    105         if (last == null)  
    106             last = first;  
    107     }  
    108     /* 
    109      * 清空缓存 
    110      */  
    111     public void clear() {  
    112         first = null;  
    113         last = null;  
    114         currentSize = 0;  
    115     }  
    116   
    117 }  
    118   
    119 class Entry {  
    120     Entry prev;//前一节点  
    121     Entry next;//后一节点  
    122     Object value;//
    123     Object key;//
    124 }  

    转自:http://blog.csdn.net/beiyeqingteng/article/details/7010411

  • 相关阅读:
    Ubuntu 10.04安装google拼音输入法
    Ubuntu 10.04 编译Android 2.1源码
    Android make sdk 错误解决方案
    关于android内核从linux内核分支上除名
    odex打包为可用的apk程序
    取得当前屏幕的截图
    android设备作为视频监控客户端的思路
    政府网站群系统选型
    浅谈网站群的一代与二代技术
    我的2013
  • 原文地址:https://www.cnblogs.com/zl1991/p/6262837.html
Copyright © 2020-2023  润新知