5月18日：top10面试算法-LRUcache的实现

问题描述：

LRU算法:优先把那些最长时间没使用的对象置换掉。根据数据的历史访问记录来进行淘汰数据，其核心思想是“如果数据最近被访问过，那么将来被访问的几率也更高”。

JAVA实现：

测试：

public class preface {

    public static void main(String[] args) {
        LRUCache cache = new LRUCache();
        cache.set(1, "one");
        cache.set(2, "two");
        cache.set(3, "three");
        cache.set(5, "five");
        cache.set(4, "five");
        cache.get(2);
        cache.set(6, "six");
        cache.printCache();// 2,3,4,5,6
    }
}

仓促之下写了个脑残版的，到头来也没能正确运行：

public class LRUCache {

    private Map<Integer, pair> cache;
    private int _cur; //计数
    private int _capacity;
    private int _nums = 0;
    private heap _h;

    public class pair implements Comparable  {

        public String value;
        public int num;

        public pair(int num, String value) {
            this.num = num;
            this.value = value;
        }

        @Override
        public int compareTo(Object o) {
            return num - ((pair) o).num;
        }
    }

    //通过堆来对pairs进行排序
    public class heap {
        public pair[] pairs;

        private int _cur;

        public heap() {
            pairs = new pair[5];
            for (int i = 0; i < 5; ++i) {
                pairs[i] = new pair(0, null);
            }
            _cur = 0;
        }

        //每次只能修改最上面的pair
        public pair set(pair p) {
            pair t = pairs[0];
            pairs[0] = p;
            return t;
        }

        public void insert(pair p) {
            if (_cur < 5) {
                pairs[_cur] = p;
            }
            ++_cur;
        }

        public void sort() {
            Arrays.sort(pairs);
        }

        public int find(pair p) {
            for (int i = 0; i < 5; ++i) {
                if (pairs[i] == p) { //==比较的是地址，equal比较的是值
                    return i;
                }
            }
            return -1;
        }
    }

    public LRUCache() {
        _cur = 0;
        _capacity = 0;
        cache = new HashMap<Integer, pair>();
        _h = new heap();
    }

    public String get(int key) {
        ++_cur;
        if (cache.containsKey(key)) {
            return cache.get(key).value;
        }
        return null;
    }

    public void set(int key, String value) {
        ++_cur;
        if (!cache.containsKey(key)) {
            ++_nums;
            pair p = new pair(_cur, value);
            if (_nums == _capacity) {
                pair t = _h.set(p);
                for (Integer k:cache.keySet()) {
                    if (cache.get(k) == t) {
                        cache.remove(t);
                        break;
                    }
                }
            } else {
                _h.insert(p);
            }
            cache.put(key, p);
        } else {
            pair p = cache.get(key);
            int t = _h.find(p);
            _h.pairs[t].num = _cur;
            _h.pairs[t].value = value;
        }
        _h.sort();
    }

    public void printCache() {
        System.out.println("print cache usage");
        for (Integer key:cache.keySet()) {
            pair p = cache.get(key);
            System.out.println("Key:" + key + " Value:" + p.value + " num:" + p.num);
        }
    }
}

C++实现：

struct DoubleLinkList {
    int key;
    int value;
    DoubleLinkList *next;
    DoubleLinkList *pre;
    DoubleLinkList(int k, int v) : key(k), value(v), next(NULL), pre(NULL) {};
};

class LRUCache{
public:
    LRUCache(int capacity) {
        size = capacity;
        cacheMap.clear();
        head = new DoubleLinkList(0, 0);
        tail = new DoubleLinkList(0, 0);
        head->next = tail;
        tail->pre = head;
    }

    int get(int key) {
        if (cacheMap.find(key) != cacheMap.end()) {
            DoubleLinkList *p = cacheMap[key];
            spliceNode(p);
            addToFront(p);
            return p->value;
        }
        return -1;
    }

    void set(int key, int value) {
        if (cacheMap.find(key) == cacheMap.end()) {
            DoubleLinkList *p = NULL;
            if (cacheMap.size() == size) {
                p = tail->pre;
                cacheMap.erase(p->key);
                p->key = key;
                p->value = value;
                spliceNode(p);
                addToFront(p);
            } else {
                p = new DoubleLinkList(key, value);
                addToFront(p);
            }
            cacheMap[key] = p;
        } else {
            DoubleLinkList *p = cacheMap[key];
            p->value = value;
            spliceNode(p);
            addToFront(p);
        }
    }

private:
    int size;
    map<int, DoubleLinkList*> cacheMap;
    DoubleLinkList *head, *tail;

    void addToFront(DoubleLinkList *p) {
        p->pre = head;
        p->next = head->next;
        head->next->pre = p;
        head->next = p;
    }

    void spliceNode(DoubleLinkList *p) {
        p->pre->next = p->next;
        p->next->pre = p->pre;
    }
};

　　用list代替双向链表：

struct CacheNode {
    int key;
    int value;
    CacheNode(int k, int v) : key(k), value(v) {
    }
};

class LRUCache{
public:
    LRUCache(int capacity) {
        size = capacity;
    }

    int get(int key) {
        if (cacheMap.find(key) != cacheMap.end()) {
            auto it = cacheMap[key];
            cacheList.splice(cacheList.begin(), cacheList, it);
            cacheMap[key] = cacheList.begin();
            return cacheList.begin()->value;
        } else {
            return -1;
        }
    }

    void set(int key, int value) {
        if (cacheMap.find(key) == cacheMap.end()) {
            if (cacheList.size() == size) {
                cacheMap.erase(cacheList.back().key);
                cacheList.pop_back();
            }
            cacheList.push_front(CacheNode(key, value));
            cacheMap[key] = cacheList.begin();
        } else {
            auto it = cacheMap[key];
            it->value = value;
            cacheList.splice(cacheList.begin(), cacheList, it);
            cacheMap[key] = cacheList.begin();
        }
    }

private:
    int size;
    list<CacheNode> cacheList;
    unordered_map<int, list<CacheNode>::iterator > cacheMap;
};