Jackson在序列化对象的时候,如果对象里面有循环依赖的情况,会报栈溢出,示例如下
@Getter
@Setter
@NoArgsConstructor
class Boss{
String name;
String department;
List<Employee> employees;
}
@Getter
@Setter
@NoArgsConstructor
class Employee{
String name;
Boss boss;
}
测试代码如下:
@Test
public void JsonManagedReferenceAndJsonBackReferenceTest() throws Exception{
CombineJacksonAnnotation.Employee employee1 = new CombineJacksonAnnotation.Employee();
employee1.setName("employee1");
CombineJacksonAnnotation.Employee employee2 = new CombineJacksonAnnotation.Employee();
employee2.setName("employee2");
CombineJacksonAnnotation.Boss boss = new CombineJacksonAnnotation.Boss();
boss.setName("boss");
boss.setDepartment("cto");
boss.setEmployees(Lists.newArrayList(employee1,employee2));
employee1.setBoss(boss);
employee2.setBoss(boss);
System.out.println(om.writeValueAsString(boss));
运行测试代码
image.png针对这种情况,Jackson提供了@JsonBackReference,加上此注解的字段,不会被序列化,也就打断了循环依赖,如下
@Getter
@Setter
@NoArgsConstructor
class Employee{
String name;
@JsonBackReference
Boss boss;
}
再次运行测试代码,结果如下
{
"name" : "boss",
"department" : "cto",
"employees" : [ {
"name" : "employee1"
}, {
"name" : "employee2"
} ]
}
可以看到,在序列化的时候,成功的将Employee里面的Boss忽略掉了,按照这样的逻辑,貌似@JsonBackReference 与@JsonIgnore很相似,我们可以试验下,将@JsonBackReference替换成@JsonIgnore,如下
@Getter
@Setter
@NoArgsConstructor
class Employee{
String name;
@JsonIgnore
Boss boss;
}
再次运行代码,可以得到相同的序列化结果,那么@JsonBackReference与@JsonIgnore的区别在哪里呢,区别主要体现在反序列的时候,我们将上面例子中序列化的结果进行反序列化,看看效果,代码如下
image.png可以看到Employee里面的Boss属性没有被赋值,但是我们将代码修改如下
@Getter
@Setter
@NoArgsConstructor
class Boss{
String name;
String department;
@JsonManagedReference
List<Employee> employees;
}
@Getter
@Setter
@NoArgsConstructor
class Employee{
String name;
@JsonBackReference
Boss boss;
}
再次进行反序列化,奇迹出现了
image.png
当属性分别打上@JsonManagedReference 与 @JsonBackReference时,Jackson会知道这两个属性间有父子关系,反序列化初始化的时候会建立起循环依赖。
可以说这一对注解是解决父子间循环依赖的利器。
@JsonIdentityInfo
@JsonIdentityInfo也可以解决父子之间的依赖关系,但是比上面介绍的两个注解更加的灵活,在上面的两个注解中,我们自己明确类之间的父子关系,但是@JsonIdentityInfo是独立的,解决的是相互之间的依赖关系,没有父子之间的上下关系。
使用方法如下
@Getter
@Setter
@NoArgsConstructor
@JsonIdentityInfo(property = "@id",generator = ObjectIdGenerators.IntSequenceGenerator.class)
class Boss{
String name;
String department;
//@JsonManagedReference
List<Employee> employees;
}
@Getter
@Setter
@NoArgsConstructor
@JsonIdentityInfo(property = "@id",generator = ObjectIdGenerators.IntSequenceGenerator.class)
class Employee{
String name;
//@JsonBackReference
Boss boss;
}
测试方法如下
@Test
public void JsonIdentityInfoTest() throws Exception{
CombineJacksonAnnotation.Employee employee1 = new CombineJacksonAnnotation.Employee();
employee1.setName("employee1");
CombineJacksonAnnotation.Employee employee2 = new CombineJacksonAnnotation.Employee();
employee2.setName("employee2");
CombineJacksonAnnotation.Boss boss = new CombineJacksonAnnotation.Boss();
boss.setName("boss");
boss.setDepartment("cto");
boss.setEmployees(Lists.newArrayList(employee1,employee2));
employee1.setBoss(boss);
employee2.setBoss(boss);
System.out.println(om.writeValueAsString(boss));
}
输出如下
{
"@id" : 1,
"name" : "boss",
"department" : "cto",
"employees" : [ {
"@id" : 2,
"name" : "employee1",
"boss" : 1
}, {
"@id" : 3,
"name" : "employee2",
"boss" : 1
} ]
}
其中@id表明此类的唯一标签名,我们可以使用类已经存在的属性名,如下
@Getter
@Setter
@NoArgsConstructor
@JsonIdentityInfo(property = "name",generator = ObjectIdGenerators.PropertyGenerator.class)
class Boss{
String name;
String department;
//@JsonManagedReference
List<Employee> employees;
}
@Getter
@Setter
@NoArgsConstructor
@JsonIdentityInfo(property = "name",generator = ObjectIdGenerators.PropertyGenerator.class)
class Employee{
String name;
//@JsonBackReference
Boss boss;
}
输出结果如下
{
"name" : "boss",
"department" : "cto",
"employees" : [ {
"name" : "employee1",
"boss" : "boss"
}, {
"name" : "employee2",
"boss" : "boss"
} ]
}
此注解的使用在解决循环解决的时候更加的灵活。
网友评论