浅谈关于java中的深浅拷贝

一.浅拷贝(shallow copy)

---

1.如何实现浅拷贝?

Object类 是所有类的直接或间接父类,Object中存在clone方法,如下

protected native Object clone() throws CloneNotSupportedException;

如果想要使一个类的对象能够调用clone方法 ,则需要实现Cloneable接口, 并重写 clone方法:

public class Student implements Cloneable{


    private int sno ;
    private String name;

    //getter ,setter 省略

    @Override
    public Object clone() throws CloneNotSupportedException {
        Student s = null;
        try{
            s = (Student)super.clone();
        }catch (Exception e){
            e.printStackTrace();
        }
        return s;
    }

}

现在测试clone方法:

@Test
    public void test04() throws CloneNotSupportedException {
        //创建Student对象
        Student s1 = new Student();
        s1.setSno(1);
        s1.setName("Rye");

        //通过clone 拷贝一个对象
        Student s2 = (Student)s1.clone();

        System.out.println("s1:"+s1);
        System.out.println("s2:"+s2);
        System.out.println("s1 == s2 ? ==> "+(s1 == s2));
    }

按照预期,克隆出的对象s2中的字段值应该与s1相同,但与s1对应的对象不在同一块内存空间,结果如下:

s1:Student{sno=1, name='Rye'}
s2:Student{sno=1, name='Rye'}
s1 == s2 ? ==> false

此时如果修改 s1中的sno为2,那么会不会影响到s2中的sno呢?

//修改s1中的sno
s1.setSno(2);

结果如下:

s1:Student{sno=2, name='Rye'}
s2:Student{sno=1, name='Rye'}

此时看似已经完成了 copy, s1 与 s2有着自己不同的值,但如果为Student中新增了Teacher类型的成员变量,结果还是跟上面一样吗?让我们改造下代码:

public class Teacher {

    private int tno;
    private String name;
  
    //getter setter省略...  
}

public class Student  implements Cloneable{

    private int sno ;
    private String name;
    private Teacher teacher;

    //getter ,setter ,toString 省略...

    @Override
    public Object clone() throws CloneNotSupportedException {
        Student s = null;
        try{
            s = (Student)super.clone();
        }catch (Exception e){
            e.printStackTrace();
        }
        return s;
    }    
}

此时测试代码如下:

    @Test
    public void test02() throws CloneNotSupportedException {
        Student student1 = new Student();
        student1.setSno(1);
        student1.setName("Rye");

        Teacher teacher = new Teacher();
        teacher.setTno(1);
        teacher.setName("LinTong");

        student1.setTeacher(teacher);

        Student student2 = (Student)student1.clone();

        System.out.println("student1:"+student1);
        System.out.println("student2:"+student2);
        System.out.println("student1 == student2 ? ==> "+ (student1 ==student2));
        System.out.println("student1.teacher == student2.teacher ? ==> "+ (student1.getTeacher() ==student2.getTeacher()));

    }

运行结果如下:

student1:Student{sno=1, name='Rye', teacher=Teacher{tno=1, name='LinTong'}}
student2:Student{sno=1, name='Rye', teacher=Teacher{tno=1, name='LinTong'}}
student1 == student2 ? ==> false
student1.teacher == student2.teacher ? ==> true

由此可见,此时经过clone生成的student2, 与 student1.二者中的teacher字段, 指向同一块内存空间;

那么可能会有人问,这会有什么影响吗? 

我们拷贝的目的,更多的时候是希望获得全新并且值相同的对象,操作原始对象或拷贝的新对象,对彼此之间互不影响;

此时我们修改student1中teacher的tno ,如下:

//修改teacher中的 tno值为2
student1.getTeacher().setTno(2);

再次运行test:

student1:Student{sno=1, name='Rye', teacher=Teacher{tno=2, name='LinTong'}}
student2:Student{sno=1, name='Rye', teacher=Teacher{tno=2, name='LinTong'}}
student1 == student2 ? ==> false
student1.teacher == student2.teacher ? ==> true

此时发现,student2中的teacher的tno ,也跟着变化了.

变化的原因是:通过student1执行clone时,基本类型会完全copy一份到student2对应对象内存空间中, 但是对于Teacher对象仅仅是copy了一份Teacher的引用而已.

而student1 与 student2的引用 指向的是同一块堆内存,因此不论是通过student1或是student2修改teacher 都会影响另外一个;

通过图会更直观一些:

2.浅拷贝中引用类型的变量拷贝的是对象的引用 , 可通过如下思路解决:

Teacher类中也覆写clone方法:

    @Override
    protected Object clone() {
        Teacher teacher = null;
        try {
            teacher = (Teacher)super.clone();
        } catch (CloneNotSupportedException e) {
            e.printStackTrace();
        }
        return teacher;
    }

修改Student中的clone方法,如下:

    @Override
    public Object clone() {
        Student s = null;
        try{
            s = (Student)super.clone();
            Teacher t = (Teacher)this.teacher.clone();
            s.setTeacher(t);
        }catch (Exception e){
            e.printStackTrace();
        }
        return s;
    }

此时再次运行test:

student1:Student{sno=1, name='Rye', teacher=Teacher{tno=2, name='LinTong'}}
student2:Student{sno=1, name='Rye', teacher=Teacher{tno=1, name='LinTong'}}
student1 == student2 ? ==> false
student1.teacher == student2.teacher ? ==> false

由此可见,在copy Student的同时 将Teacher也进行了修改,如图:

目前来看是满足了我们的需求,但是如果Teacher类中,同样也有别的引用类型 的成员变量呢?

那么就同样需要一直覆写clone方法,如果这个关系不是特多还可以接受,如果引用关系很复杂就会显得代码繁琐;

此时应该使用序列化完成深度拷贝;

二.深拷贝(deep copy)

---

使用序列化完成深拷贝

深拷贝是利用对象流,将对象序列化,再反序列化得出新的对象. 因此首先需要实现序列化接口,如下:

public class Student implements Serializable{

    private static final long serialVersionUID = -2232725257771333130L;

    private int sno ;
    private String name;
    private Teacher teacher;
　　//getter ,setter,toString()省略...
}

Teacher也要实现序列化接口:

public class Teacher implements Serializable{
    private static final long serialVersionUID = 4477679176385287943L;
    private int tno;
    private String name;
　　

　//getter ,setter,toString()省略...

}

工具方法:

　　//工具方法
　　public Object cloneObject(Object object) throws IOException, ClassNotFoundException {
        //将对象序列化
        ByteArrayOutputStream outputStream = new ByteArrayOutputStream();
        ObjectOutputStream objectOutputStream = new ObjectOutputStream(outputStream);
        objectOutputStream.writeObject(object);

        //将字节反序列化
        ByteArrayInputStream inputStream = new ByteArrayInputStream(outputStream.toByteArray());
        ObjectInputStream objectInputStream = new ObjectInputStream(inputStream);
        Object obj = objectInputStream.readObject();

        return obj;
    }

测试类:

　　 public void test05() throws IOException, ClassNotFoundException {
        Student student1 = new Student();
        student1.setSno(1);
        student1.setName("Rye");

        Teacher teacher = new Teacher();
        teacher.setTno(1);
        teacher.setName("LinTong");

        student1.setTeacher(teacher);

        Student student2 = (Student)cloneObject(student1);
        //修改teacher中的 tno值为2
        student1.getTeacher().setTno(2);

        System.out.println("student1:"+student1);
        System.out.println("student2:"+student2);
        System.out.println("student1 == student2 ? ==> "+ (student1 ==student2));
        System.out.println("student1.teacher == student2.teacher ? ==> "+ (student1.getTeacher() ==student2.getTeacher()));
    }

如果Teacher类或者Student类没有实现序列化接口,则执行时会报异常,如下:

java.io.NotSerializableException: com.example.test.Teacher

在都实现了Serializable接口的情况下,运行结果如下:

student1:Student{sno=1, name='Rye', teacher=Teacher{tno=2, name='LinTong'}}
student2:Student{sno=1, name='Rye', teacher=Teacher{tno=1, name='LinTong'}}
student1 == student2 ? ==> false
student1.teacher == student2.teacher ? ==> false

由此通过对象流的方式,成功完成了深度拷贝;

---

三.重写clone方法 与 通过序列化 两种拷贝方式比较:

clone方法:

优点:速度快,效率高

缺点:在对象引用比较深时,使用此方式比较繁琐

通过序列化:

优点:非常简便的就可以完成深度copy

缺点:由于序列化的过程需要跟磁盘打交道,因此效率会低于clone方式

如何抉择?

实际开发中,根据两种方式的优缺点进行选择即可!