跳跃表的原理及Java实现

原理参考：SkipList跳表

这里我使用Java实现其原理：

首先是SkipListNode的定义：

SkipListNode.java

package skiplist;

/**
 * Created by zhuxinquan on 17-3-11.
 */
public class SkipListNode implements Comparable {

    private int value;
    private SkipListNode next = null;
    private SkipListNode downNext = null;

    @Override
    protected void finalize() throws Throwable {
        super.finalize();
        System.out.printf("123");
    }

    public int getValue() {
        return value;
    }

    public void setValue(int value) {
        this.value = value;
    }

    public SkipListNode getNext() {
        return next;
    }

    public void setNext(SkipListNode next) {
        this.next = next;
    }

    public SkipListNode getDownNext() {
        return downNext;
    }

    public void setDownNext(SkipListNode downNext) {
        this.downNext = downNext;
    }

    @Override
    public int compareTo(Object o) {
        return this.value > ((SkipListNode)o).value ? 1 : -1;
    }
}

然后是跳跃表的具体操作，操作过程在开始部分处的链接已明了，这里直接模拟实现：

SkipList.java

package skiplist;

import java.util.Random;

/**
 * Created by zhuxinquan on 17-3-11.
 */
public class SkipList {

    public int level = 0;
    public SkipListNode top = null;

    public SkipList() {
        this(4);
    }

    //跳跃表的初始化
    public SkipList(int level) {
        this.level = level;
        SkipListNode skipListNode = null;
        SkipListNode temp = top;
        SkipListNode tempDown = null;
        SkipListNode tempNextDown = null;
        int tempLevel = level;
        while(tempLevel -- != 0){
            skipListNode = createNode(Integer.MIN_VALUE);
            temp = skipListNode;
            skipListNode = createNode(Integer.MAX_VALUE);
            temp.setNext(skipListNode);
            temp.setDownNext(tempDown);
            temp.getNext().setDownNext(tempNextDown);
            tempDown = temp;
            tempNextDown = temp.getNext();
        }
        top = temp;
    }

    //随机产生数k，k层下的都需要将值插入
    public int randomLevel(){
        int k = 1;
        while(new Random().nextInt()%2 == 0){
            k ++;
        }
        return k > level ? level : k;
    }

    //查找
    public SkipListNode find(int value){
        SkipListNode node = top;
        while(true){
            while(node.getNext().getValue() < value){
                node = node.getNext();
            }
            if(node.getDownNext() == null){
                //返回要查找的节点的前一个节点
                return node;
            }
            node = node.getDownNext();
        }
    }

    //删除一个节点
    public boolean delete(int value){
        int tempLevel = level;
        SkipListNode skipListNode = top;
        SkipListNode temp = skipListNode;
        boolean flag = false;
        while(tempLevel -- != 0){
            while(temp.getNext().getValue() < value){
                temp = temp.getNext();
            }
            if(temp.getNext().getValue() == value){
                temp.setNext(temp.getNext().getNext());
                flag = true;
            }
            temp = skipListNode.getDownNext();
        }
        return flag;
    }

    //插入一个节点
    public void insert(int value){
        SkipListNode skipListNode = null;
        int k = randomLevel();
        SkipListNode temp = top;
        int tempLevel = level;
        SkipListNode tempNode = null;
        //当在第n行插入后，在第n - 1行插入时需要将第n行backTempNode的DownNext域指向第n - 1的域
        SkipListNode backTempNode = null;
        int flag = 1;
        while(tempLevel-- != k){
            temp = temp.getDownNext();
        }

        tempLevel++;
        tempNode = temp;
        //小于k层的都需要进行插入
        while(tempLevel-- != 0){
            //在第k层寻找要插入的位置
            while(tempNode.getNext().getValue() < value){
                tempNode = tempNode.getNext();
            }
            skipListNode = createNode(value);
            //如果是顶层
            if(flag != 1){
                backTempNode.setDownNext(skipListNode);
            }
            backTempNode = skipListNode;
            skipListNode.setNext(tempNode.getNext());
            tempNode.setNext(skipListNode);
            flag = 0;
            tempNode = tempNode.getDownNext();
        }
    }

    //创建一个节点
    private SkipListNode createNode(int value){
        SkipListNode node =  new SkipListNode();
        node.setValue(value);
        return node;
    }
}

然后再定义一个测试类如下：

SkipListTest.java

package skiplist;

import java.util.HashSet;
import java.util.LinkedList;
import java.util.List;
import java.util.Set;

/**
 * Created by zhuxinquan on 17-3-11.
 */
public class SkipListTest {
    public static void main(String[] args) {
        long startTime = 0;
        long endTime = 0;
        SkipList skipList = new SkipList(12);
        Set<Integer> set = new HashSet();
        List<Integer> list = new LinkedList<>();

        //测试跳跃表性能
        startTime = System.currentTimeMillis();
        for (int i = 1; i < 100000; i++) {
            skipList.insert(i);
        }
        endTime = System.currentTimeMillis();
        System.out.printf("createSkipList:%d\n", endTime - startTime);
        startTime = System.currentTimeMillis();
        System.out.printf("find(555)：%d\n", skipList.find(55555).getNext().getValue());
        endTime = System.currentTimeMillis();
        System.out.printf("skipListFindTime:%d\n\n", endTime - startTime);

        //测试LinkedList性能
        startTime = System.currentTimeMillis();
        for (int i = 1; i < 100000; i++) {
            list.add(i);
        }
        endTime = System.currentTimeMillis();
        System.out.printf("createList:%d\n", endTime - startTime);
        startTime = System.currentTimeMillis();
        System.out.printf("find(555)：%b\n", list.contains(55555));
        endTime = System.currentTimeMillis();
        System.out.printf("linkedListFindTime:%d\n\n", endTime - startTime);

        //测试hashSet性能
        startTime = System.currentTimeMillis();
        for (int i = 1; i < 100000; i++) {
            set.add(i);
        }
        endTime = System.currentTimeMillis();
        System.out.printf("createSet:%d\n", endTime - startTime);
        startTime = System.currentTimeMillis();
        System.out.printf("find(555)：%b\n", set.contains(55555));
        endTime = System.currentTimeMillis();
        System.out.printf("hashSetFindTime:%d\n", endTime - startTime);
    }
}

 

上述测试类对100000条数据进行插入和查找，分别统计使用跳跃表、LinkedList、HashSet时插入和查找的时间，结果如下：

下面是50000整数中查找38555这个值消耗的时间：

上图很明显的可以看出，跳跃表的查找效率还是很高的（在数据量高的时候），同时，跳跃表也肯定消耗了不少的额外空间。

上述代码只是个人根据自己的理解简单实现了跳跃表的原理，其中自然有不少瑕疵，也有好多可优化的地方，比如可以将每层的起点节点的引用保存在一个一维数组当中、跳表中存储键值对这样的数据结构等等