Java中的深拷贝(深复制)和浅拷贝(浅复制)

深拷贝(深复制)和浅拷贝(浅复制)是两个比较通用的概念,尤其在C++语言中,若不弄懂,则会在delete的时候出问题,但是我们在这幸好用的是Java。虽然java自动管理对象的回收,但对于深拷贝(深复制)和浅拷贝(浅复制),我们还是要给予足够的重视,因为有时这两个概念往往会给我们带来不小的困惑。

浅拷贝是指拷贝对象时仅仅拷贝对象本身(包括对象中的基本变量),而不拷贝对象包含的引用指向的对象。深拷贝不仅拷贝对象本身,而且拷贝对象包含的引用指向的所有对象。举例来说更加清楚:对象A1中包含对B1的引用,B1中包含对C1的引用。浅拷贝A1得到A2,A2 中依然包含对B1的引用,B1中依然包含对C1的引用。深拷贝则是对浅拷贝的递归,深拷贝A1得到A2,A2中包含对B2(B1的copy)的引用,B2 中包含对C2(C1的copy)的引用。

若不对clone()方法进行改写,则调用此方法得到的对象即为浅拷贝,下面我们着重谈一下深拷贝。

运行下面的程序,看一看浅拷贝:

 

class Professor0 implements Cloneable {
    String name;
    int age;
 
    Professor0(String name, int age) {
        this.name = name;
        this.age = age;
    }
 
    public Object clone() throws CloneNotSupportedException {
        return super.clone();
    }
}
 
class Student0 implements Cloneable {
    String name;// 常量对象。
    int age;
    Professor0 p;// 学生1和学生2的引用值都是一样的。
 
    Student0(String name, int age, Professor0 p) {
        this.name = name;
        this.age = age;
        this.p = p;
    }
 
    public Object clone() {
        Student0 o = null;
        try {
            o = (Student0) super.clone();
        } catch (CloneNotSupportedException e) {
            System.out.println(e.toString());
        }
 
        return o;
    }
}
 
public class ShallowCopy {
    public static void main(String[] args) {
        Professor0 p = new Professor0("wangwu", 50);
        Student0 s1 = new Student0("zhangsan", 18, p);
        Student0 s2 = (Student0) s1.clone();
        s2.p.name = "lisi";
        s2.p.age = 30;
        s2.name = "z";
        s2.age = 45;
        System.out.println("学生s1的姓名:" + s1.name + "\n学生s1教授的姓名:" + s1.p.name + "," + "\n学生s1教授的年纪" + s1.p.age);// 学生1的教授
    }
}

 

s2变了,但s1也变了,证明s1的p和s2的p指向的是同一个对象。这在我们有的实际需求中,却不是这样,因而我们需要深拷贝:

 

class Professor implements Cloneable {
    String name;
    int age;
 
    Professor(String name, int age) {
        this.name = name;
        this.age = age;
    }
 
    public Object clone() {
        Object o = null;
        try {
            o = super.clone();
        } catch (CloneNotSupportedException e) {
            System.out.println(e.toString());
        }
        return o;
    }
}
 
class Student implements Cloneable {
    String name;
    int age;
    Professor p;
 
    Student(String name, int age, Professor p) {
        this.name = name;
        this.age = age;
        this.p = p;
    }
 
    public Object clone() {
        Student o = null;
        try {
            o = (Student) super.clone();
        } catch (CloneNotSupportedException e) {
            System.out.println(e.toString());
        }
        o.p = (Professor) p.clone();
        return o;
    }
}
 
public class DeepCopy {
    public static void main(String args[]) {
        long t1 = System.currentTimeMillis();
        Professor p = new Professor("wangwu", 50);
        Student s1 = new Student("zhangsan", 18, p);
        Student s2 = (Student) s1.clone();
        s2.p.name = "lisi";
        s2.p.age = 30;
        System.out.println("name=" + s1.p.name + "," + "age=" + s1.p.age);// 学生1的教授不改变。
        long t2 = System.currentTimeMillis();
        System.out.println(t2-t1);
    }
}

 

当然我们还有一种深拷贝方法,就是将对象串行化:

 

import java.io.*;
//Serialization is time-consuming
class Professor2 implements Serializable {
    /**
     *
     */
    private static final long serialVersionUID = 1L;
    String name;
    int age;
 
    Professor2(String name, int age) {
        this.name = name;
        this.age = age;
    }
}
 
class Student2 implements Serializable {
    /**
     *
     */
    private static final long serialVersionUID = 1L;
    String name;// 常量对象。
    int age;
    Professor2 p;// 学生1和学生2的引用值都是一样的。
 
    Student2(String name, int age, Professor2 p) {
        this.name = name;
        this.age = age;
        this.p = p;
    }
 
    public Object deepClone() throws IOException, OptionalDataException,
            ClassNotFoundException {
        // 将对象写到流里
        ByteArrayOutputStream bo = new ByteArrayOutputStream();
        ObjectOutputStream oo = new ObjectOutputStream(bo);
        oo.writeObject(this);
        // 从流里读出来
        ByteArrayInputStream bi = new ByteArrayInputStream(bo.toByteArray());
        ObjectInputStream oi = new ObjectInputStream(bi);
        return (oi.readObject());
    }
 
}
 
public class DeepCopy2 {
 
    /**
     * @param args
     */
    public static void main(String[] args) throws OptionalDataException,
            IOException, ClassNotFoundException {
        long t1 = System.currentTimeMillis();
        Professor2 p = new Professor2("wangwu", 50);
        Student2 s1 = new Student2("zhangsan", 18, p);
        Student2 s2 = (Student2) s1.deepClone();
        s2.p.name = "lisi";
        s2.p.age = 30;
        System.out.println("name=" + s1.p.name + "," + "age=" + s1.p.age); // 学生1的教授不改变。
        long t2 = System.currentTimeMillis();
        System.out.println(t2-t1);
    }
 
}

 

但是串行化却很耗时,在一些框架中,我们便可以感受到,它们往往将对象进行串行化后进行传递,耗时较多。

posted @ 2017-05-12 10:43  Mason.Ke  阅读(229)  评论(0编辑  收藏  举报