如何在Java中深拷贝ArrayList
大约 4 分钟
如何在Java中深拷贝ArrayList
1. 引言
在这篇简短的教程中，我们将学习如何在Java中复制一个_ArrayList_，重点介绍创建列表中元素的深拷贝的不同方法。
2. 浅拷贝与深拷贝
浅拷贝技术复制原始对象，但只复制可变字段的引用，而不是实际对象。另一方面，深拷贝创建了所有可变字段的独立副本，包括深度嵌套的对象。有关详细指南，请参阅我们的文章《深拷贝和浅拷贝之间的区别》。
3. 模型
让我们创建两个类：Course_和_Student。_Student_类有一个_Course_对象的实例作为可变的依赖项：
public class Course {
private Integer courseId;
private String courseName;
// 标准 getter 和 setter
}
public class Student {
private int studentId;
private String studentName;
private Course course;
// 标准 getter 和 setter
}
4. 使用_Cloneable_接口进行深拷贝
让我们实现标记接口_Cloneable_并在我们的模型类中重写_clone_方法以创建深拷贝：
@Override
public Course clone() {
try {
return (Course) super.clone();
} catch (CloneNotSupportedException e) {
throw new IllegalStateException(e);
}
}
注意_super.clone()_总是返回对象的浅拷贝。在_Course_类中，我们没有任何可变字段，而在_Student_类中，我们需要显式设置可变字段以创建深拷贝：
@Override
public Student clone() {
Student student;
try {
student = (Student) super.clone();
} catch (CloneNotSupportedException e) {
throw new IllegalStateException(e);
}
student.course = this.course.clone();
return student;
}
现在，让我们遍历项目并使用_clone_方法，并验证是否创建了深拷贝：
public static List````````````````<Student>```````````````` deepCopyUsingCloneable(List````````````````<Student>```````````````` students){
return students.stream().map(Student::clone).collect(Collectors.toList());
}
@Test
public void whenCreatingCopyWithCloneable_thenObjectsShouldNotBeSame() {
Course course = new Course(1, "Spring Masterclass");
Student student1 = new Student(1, "John", course);
Student student2 = new Student(2, "David", course);
List````````````````<Student>```````````````` students = new ArrayList<>();
students.add(student1);
students.add(student2);
List````````````````<Student>```````````````` deepCopy = Student.deepCopyUsingCloneable(students);
Assertions.assertEquals(students.get(0), deepCopy.get(0));
Assertions.assertNotSame(students.get(0),deepCopy.get(0));
Assertions.assertEquals(students.get(1), deepCopy.get(1));
Assertions.assertNotSame(students.get(1),deepCopy.get(1));
}
5. 使用复制构造函数进行深拷贝
复制构造函数是一个特殊的构造函数，它接受其类类型的参数，并回一个具有传递值的新类实例。
让我们为_Student_对象创建一个复制构造函数，并使用它来对列表中的每个项目进行深拷贝：
public Student(Student student) {
this.studentId = student.getStudentId();
this.studentName = student.getStudentName();
this.course = new Course(student.getCourse()
.getCourseId(), student.getCourse()
.getCourseName());
}
接下来，让我们遍历列表中的项目，并使用上面创建的复制构造函数对列表中的每个项目进行深拷贝，并返回一个新列表：
public static List````````````````<Student>```````````````` deepCopyUsingCopyConstructor(List````````````````<Student>```````````````` students){
return students.stream().map(Student::new).collect(Collectors.toList());
}
在这种情况下，修改原始_ArrayList_或列表中的元素对复制的列表没有任何影响，反之亦然：
@Test
public void whenCreatingDeepCopyWithCopyConstructor_thenObjectsShouldNotBeSame() {
Course course = new Course(1, "Spring Masterclass");
Student student1 = new Student(1, "John", course);
Student student2 = new Student(2, "David", course);
List````````````````<Student>```````````````` students = new ArrayList<>();
students.add(student1);
students.add(student2);
List````````````````<Student>```````````````` deepCopy = Student.deepCopyUsingCopyConstructor(students);
Assertions.assertEquals(students.get(0), deepCopy.get(0));
Assertions.assertNotSame(students.get(0),deepCopy.get(0));
Assertions.assertEquals(students.get(1), deepCopy.get(1));
Assertions.assertNotSame(students.get(1),deepCopy.get(1));
}
6. 使用Apache Commons库进行深拷贝
Apache Commons库提供了一个实用方法_SerializationUtils.clone()_，它通过序列化和反序列化帮助制作对象的深拷贝。有关序列化的详细指南，请参阅我们的文章Java序列化。
使用这种方法确保了对象图中的所有字段，包括深度嵌套的对象都被复制，从而得到一个完全独立的深拷贝：
public static List````````````````<Student>```````````````` deepCopyUsingSerialization(List````````````````<Student>```````````````` students){
return students.stream().map(SerializationUtils::clone).collect(Collectors.toList());
}
对象图中的所有对象必须实现_Serializable_接口才能成功。否则，将抛出异常：
@Test
public void whenCreatingDeepCopyWithSerializationUtils_thenObjectsShouldNotBeSame() {
Course course = new Course(1, "Spring Masterclass");
Student student1 = new Student(1, "John", course);
Student student2 = new Student(2, "David", course);
List````````````````<Student>```````````````` students = new ArrayList<>();
students.add(student1);
students.add(student2);
List````````````````<Student>```````````````` deepCopy = Student.deepCopyUsingSerialization(students);
Assertions.assertEquals(students.get(0), deepCopy.get(0));
Assertions.assertNotSame(students.get(0),deepCopy.get(0));
Assertions.assertEquals(students.get(1), deepCopy.get(1));
Assertions.assertNotSame(students.get(1),deepCopy.get(1));
}
这使我们免于为复杂的对象结构创建克隆逻辑。然而，由于进行序列化和反序列化的开销，它比其他方法稍微慢一些。
Apache Commons lang3库的最新版本可以在Maven中央仓库中找到。
7. 使用Jackson库进行深拷贝
Jackson是另一个使用序列化和反序列化创建原始对象深拷贝的库。它将对象序列化为JSON字符串，然后将其反序列化回一个新的独立副本：
public static Student createDeepCopy(Student student) {
ObjectMapper objectMapper = new ObjectMapper();
try {
return objectMapper.readValue(objectMapper.writeValueAsString(student), Student.class);
} catch (JsonProcessingException e) {
throw new IllegalArgumentException(e);
}
}
public static List````````````````<Student>```````````````` deepCopyUsingJackson(List````````````````<Student>```````````````` students) {
return students.stream().map(Student::createDeepCopy).collect(Collectors.toList());
}
请注意，Jackson需要存在一个默认构造函数才能序列化和反序列化任何给定的对象：
@Test
public void whenCreatingDeepCopyWithJackson_thenObjectsShouldNotBeSame() {
Course course = new Course(1, "Spring Masterclass");
Student student1 = new Student(1, "John", course);
Student student2 = new Student(2, "David", course);
List````````````````<Student>```````````````` students = new ArrayList<>();
students.add(student1);
students.add(student2);
List````````````````<Student>```````````````` deepCopy = Student.deepCopyUsingJackson(students);
Assertions.assertEquals(students.get(0), deepCopy.get(0));
Assertions.assertNotSame(students.get(0),deepCopy.get(0));
Assertions.assertEquals(students.get(1), deepCopy.get(1));
Assertions.assertNotSame(students.get(1),deepCopy.get(1));
}
Jackson-databind库的最新版本可以在Maven中央仓库中找到。
8. 结论
在本教程中，我们涵盖了复制_ArrayList_的各种方式，包括原生方法和使用第三方库。如常，源代码可在GitHub上找到。