Skip to content

Latest commit

 

History

History
208 lines (166 loc) · 5.63 KB

146 LRU缓存机制.md

File metadata and controls

208 lines (166 loc) · 5.63 KB

题目

运用你所掌握的数据结构,设计和实现一个 LRU (最近最少使用) 缓存机制。它应该支持以下操作: 获取数据 get 和 写入数据 put 。

获取数据 get(key) - 如果密钥 (key) 存在于缓存中,则获取密钥的值(总是正数),否则返回 -1。

写入数据 put(key, value) - 如果密钥不存在,则写入其数据值。当缓存容量达到上限时,它应该在写入新数据之前删除最近最少使用的数据值,从而为新的数据值留出空间。

进阶:

你是否可以在 O(1) 时间复杂度内完成这两种操作?

示例:

LRUCache cache = new LRUCache( 2 /* 缓存容量 */ );

cache.put(1, 1);
cache.put(2, 2);
cache.get(1);       // 返回  1
cache.put(3, 3);    // 该操作会使得密钥 2 作废
cache.get(2);       // 返回 -1 (未找到)
cache.put(4, 4);    // 该操作会使得密钥 1 作废
cache.get(1);       // 返回 -1 (未找到)
cache.get(3);       // 返回  3
cache.get(4);       // 返回  4

实现

计算机的缓存容量有限,如果缓存满了就要删除一些内容,给新内容腾位置。但问题是,删除哪些内容呢?我们肯定希望删掉哪些没什么用的缓存,而把有用的数据继续留在缓存里,方便之后继续使用。那么,什么样的数据,我们判定为「有用的」的数据呢?

LRU 缓存淘汰算法就是一种常用策略。LRU 的全称是 Least Recently Used,也就是说我们认为最近使用过的数据应该是是「有用的」,很久都没用过的数据应该是无用的,内存满了就优先删那些很久没用过的数据。

第一种:利用单链表的方式

  • 状态:通过
  • 18 / 18 个通过测试用例
  • 执行用时: 868 ms, 在所有 C# 提交中击败了 6.25% 的用户
  • 内存消耗: 47.8 MB, 在所有 C# 提交中击败了 26.67% 的用户
public class LRUCache
{
    private readonly int _length;
    private readonly List<KeyValuePair<int, int>> _lst;

    public LRUCache(int capacity)
    {
        _length = capacity;
        _lst = new List<KeyValuePair<int, int>>();
    }

    private int GetIndex(int key)
    {
        for (int i=0,len=_lst.Count;i<len;i++)
        {
            if (_lst[i].Key == key)
            {
                return i;
            }
        }
        return -1;
    }

    public int Get(int key)
    {
        int index = GetIndex(key);
        if (index!=-1)
        {
            int val = _lst[index].Value;
            _lst.RemoveAt(index);
            _lst.Add(new KeyValuePair<int, int>(key, val));
            return val;
        }
        return -1;
    }

    public void Put(int key, int value)
    {
        int index = GetIndex(key);
        if (index!=-1)
        {
            _lst.RemoveAt(index);
        }
        else if (_lst.Count == _length)
        {
            _lst.RemoveAt(0);
        }
        _lst.Add(new KeyValuePair<int, int>(key, value));
    }
}

/**
 * Your LRUCache object will be instantiated and called as such:
 * LRUCache obj = new LRUCache(capacity);
 * int param_1 = obj.Get(key);
 * obj.Put(key,value);
 */

第二种:利用 字典 + 列表 的方式

把字典当作一个存储容器,由于字典是无序的,即 dict 内部存放的顺序和 key 放入的顺序是没有关系的,所以需要一个 list 来辅助排序。

C# 语言

  • 状态:通过
  • 18 / 18 个通过测试用例
  • 执行用时: 392 ms, 在所有 C# 提交中击败了 76.56% 的用户
  • 内存消耗: 47.9 MB, 在所有 C# 提交中击败了 20.00% 的用户
public class LRUCache
{
    private readonly List<int> _keys;
    private readonly Dictionary<int, int> _dict;


    public LRUCache(int capacity)
    {
        _keys = new List<int>(capacity);
        _dict = new Dictionary<int, int>(capacity);
    }

    public int Get(int key)
    {
        if (_dict.ContainsKey(key))
        {
            _keys.Remove(key);
            _keys.Add(key);
            return _dict[key];
        }
        return -1;
    }

    public void Put(int key, int value)
    {
        if (_dict.ContainsKey(key))
        {
            _dict.Remove(key);
            _keys.Remove(key);
        }
        else if (_keys.Count == _keys.Capacity)
        {
            _dict.Remove(_keys[0]);
            _keys.RemoveAt(0);
        }
        _keys.Add(key);
        _dict.Add(key, value);
    }
}

/**
 * Your LRUCache object will be instantiated and called as such:
 * LRUCache obj = new LRUCache(capacity);
 * int param_1 = obj.Get(key);
 * obj.Put(key,value);
 */

Python 语言

  • 执行结果:通过
  • 执行用时:628 ms, 在所有 Python3 提交中击败了 12.15% 的用户
  • 内存消耗:22 MB, 在所有 Python3 提交中击败了 65.38% 的用户
class LRUCache:

    def __init__(self, capacity: int):
        self._capacity = capacity
        self._dict = dict()
        self._keys = list()

    def get(self, key: int) -> int:
        if key in self._dict:
            self._keys.remove(key)
            self._keys.append(key)
            return self._dict[key]
        return -1

    def put(self, key: int, value: int) -> None:
        if key in self._dict:
            self._dict.pop(key)
            self._keys.remove(key)
        elif len(self._keys) == self._capacity:
            self._dict.pop(self._keys[0])
            self._keys.remove(self._keys[0])
        self._keys.append(key)
        self._dict[key] = value

# Your LRUCache object will be instantiated and called as such:
# obj = LRUCache(capacity)
# param_1 = obj.get(key)
# obj.put(key,value)