Gson?So easy.

1.概述

这篇文章主要讲述了Gson的使用.包括从最基础的基本类型的序列化,到对象,数组,集合,再到Gson注解,Gson Builder,再到格式化,自定义序列化与反序列化等内容.
另外文章篇幅较长,建议挑选所需部分查看.所有例子都提供了完整源码,在文章的后面.

2.Gson是什么?

(1)JSON

JSON全称为JavaScript Object Notation,一种轻量级的数据交换格式.

类似于XML但比XML更小,更易解析.

(2)Gson

Gson是Google提供的可以使Java对象与JSON互转的类库,可将Java对象转换为JSON,也可将JSON转换成Java对象.

(3)Gson的好处

• a.容易,高效,强大:Gson是Google管理的标准化库,经过高度优化,同时api简单,比如fromJSON(),toJSON().
• b.无依赖性:不需要其他库,当然jdk除外.
• c.结果简单:转换成的json易于阅读.
• d.支持泛型,支持内部类.
• e.开源,免费提供.

3.配置Gson环境

目前最新的是2.8.6版本.

(1)Gradle

dependencies {
  implementation 'com.google.code.gson:gson:2.8.6'
}

(2)Maven

<dependency>
  <groupId>com.google.code.gson</groupId>
  <artifactId>gson</artifactId>
  <version>2.8.6</version>
</dependency>

(3)jar

点击下载.

4.Gson基础

(1)基础类型

int intJSON = 1;
long longJSON = 1L;
double doubleJSON = 1.0;
float floatJSON = 1.0f;
byte byteJSON = 1;
char charJSON = 'G';
short shortJSON = 1;
boolean boooleanJSON = true;
System.out.println(gson.toJson(intJSON));
System.out.println(gson.toJson(longJSON));
System.out.println(gson.toJson(doubleJSON));
System.out.println(gson.toJson(floatJSON));
System.out.println(gson.toJson(byteJSON));
System.out.println(gson.toJson(charJSON));
System.out.println(gson.toJson(shortJSON));
System.out.println(gson.toJson(boooleanJSON));

System.out.println("----------------------------------------");

System.out.println(gson.fromJson("1",Integer.class));
System.out.println(gson.fromJson("1.0",Double.class));
System.out.println(gson.fromJson("1",Long.class));
System.out.println(gson.fromJson("true",Boolean.class));
System.out.println(gson.fromJson("B",Character.class));

就是对应输出,没啥好说的.

(2)嵌套对象

Gson gson = new Gson();
System.out.println(gson.toJson(new A()));

String str = "{'field':'gggg','field2':33,'field3':'G','field4':'true'}";
A a = gson.fromJson(str,A.class);

System.out.println(a.getField1());
System.out.println(a.getField2());
System.out.println(a.getField3());
System.out.println(a.getField4());

A类:

class A
{
    private String field1 = "123";
    private int field2 = 1;
    private char field3 = 'X';
    private Boolean field4 = true;
}

序列化出来的对象用{}表示.

反序列化时,注意格式,注意名字对应,用单引号引起来,还有char会自动变为String类型,另外对于布尔类型可以加单引号或不加单引号,都可以正常反序列化.

(3)数组

a.普通数组

int [] a = new int []{1,2,3};
double [] b = new double []{1.0,2.0,3.0};
String [] c = new String []{"123","456"};

System.out.println(gson.toJson(a));
System.out.println(gson.toJson(b));
System.out.println(gson.toJson(c));

int [] aa = gson.fromJson("[6,7,8]",int [].class);
double [] bb = gson.fromJson("[6.0,8.0,9.0]",double [].class);
String [] cc = gson.fromJson("['123123','5464']",String [].class);

System.out.println(Arrays.toString(aa));
System.out.println(Arrays.toString(bb));
System.out.println(Arrays.toString(cc));

对普通数组的话,{}变成了[].
还有就是反序列化时,默认会在逗号后面添一个空格.

b.List

List<String> list = new ArrayList<>();
list.add("123");
list.add("456");

System.out.println(gson.toJson(list));

List<AAAA> list2 = new ArrayList<>();
list2.add(new AAAA());
list2.add(new AAAA("45345",8888,false));

System.out.println(gson.toJson(list2));
System.out.println("---------------------------------");

Type type = new TypeToken<List<String>>(){}.getType();

String strList = "['234234','45457']";
List<String> list3 = gson.fromJson(strList, type);
System.out.println(list3);

String strList2 = "[" +
    "{'field1':'9999','field2':666,'field3':'true'}" +
    "," +
    "{'field1':'sdlkfkl','field2':-234234,'field3':'false'}" +
    "]";
List<AAAA> list4 = gson.fromJson(strList2, new TypeToken<List<AAAA>>(){}.getType());
System.out.println(list4);

class AAAA 
{
    private String field1 = "123";
    private int field2 = 5;
    private boolean field3 = true;
}

序列化没什么问题,直接toJson即可,反序列化时,需要配合java.lang.reflect.Type使用,明确指出所要转换的类型:

Type type = new TypeToken<List<String>>(){}.getType();

TypeToken<T>中T为所需要的类型,再把这个Type对象传递给fromJson即可完成转换.
List序列化出来的用[]表示.

(4)Map

Map<String,Integer> map = new HashMap<>();
map.put("123",88);
map.put("2349",999);

System.out.println(gson.toJson(map));

String str = "{'123':23423423,'9999':-234234}";
Map<String,Integer> map2 = gson.fromJson(str, new TypeToken<Map<String,Integer>>(){}.getType());
System.out.println(map2.get("123").toString());
System.out.println(map2.get("9999").toString());

Map同样序列化时直接toJson,反序列化时使用java.lang.reflect.Type.Map序列化出来的用{}表示.

(5)Set

Set<String> set = new HashSet<>();
set.add("123123");
set.add("2349594");
set.add("-234()@#$@#");
System.out.println(gson.toJson(set));

String str = "['38483','@*#$(@#$)','SD<FGDF>G']";
Set<String> set2 = gson.fromJson(str, new TypeToken<Set<String>>(){}.getType());
set2.stream().forEach(System.out::println);

set序列化出来的用[]表示.

(6)null

System.out.println(gson.toJson(new AAA()));
String str = "{'field2':333,'field3':null}";
System.out.println(gson.fromJson(str, AAA.class));

class AAA
{
    private String field1 = null;
    private Integer field2 = null;
    private Double field3 = 3.0;

    @Override
    public String toString()
    {
        return "field1:"+field1+",field2:"+field2+",field3:"+field3;
    }
}

Gson会忽略空值,在序列化时看不到null对应的键值对,反序列化时,直接对应为空.

(7)混合

一个对象里面包含了List,Map,Set,null.

System.out.println(gson.toJson(new Test()));

class Test
{
    private List<String> listField = new ArrayList<>();
    private Map<String,Double> mapField = new HashMap<>();
    private Set<Integer> setField = new HashSet<>();

    private List<User> userListFiled = new ArrayList<>();
    private Map<String,User> userMapField = new HashMap<>();
    private Set<User> userSetField = new HashSet<>();

    private Set<String> nullSetField = null;
    private Map<String,User> nullMapField = null;
    private List<User> nullListField = null;

    private Long longField = 23423423423L;
    private int intField = 234234;
    private Double doubleField = 234234.23423;
    private User userField = new User();

    {
        listField.add("234234");
        mapField.put("23432",-234.0);
        setField.add(-23423);

        userListFiled.add(new User("3459",-23423,new int []{4,5,5}));
        userListFiled.add(new User());
        userMapField.put("()",new User());
        userMapField.put("------", new User("345345",3434,new int []{4,44}));
        userSetField.add(new User());
        userSetField.add(new User());
    }
}

class User
{
    private String name = "noName";
    private int age = 0;
    private int [] nums = new int[]{2,3,4};

    public User()
    {

    }

    public User(String name,int age,int [] nums)
    {
        this.name = name;
        this.age = age;
        this.nums = nums;
    }
}

这个就不反序列化了,可以看到对象,Map用{},List,Set用[].与上面的一致.

5.Gson注解

(1)@SerializedName

SerializedName注解有两个参数,分别是value与alternate.SerializedName由注解名字可以知道与序列化成的Json名字有关.

默认情况下,json的键名与对象的字段名一致,@SerializedName可以解决序列化/反序列化时json键名与对象字段名不一致的问题,使其将json可以正确映射到对应的字段.

a.单独使用value

单独使用一个参数时,即@SerializedName("xxx")或@SerializedName(value="xxx"),序列化时,对象字段会变为@SerializedName中的值,反序列化时,若不是@SerializedName()中的值则不会反序列化:

public class SerializedNameValueTest
{
    public static void main(String[] args) {
        Gson gson = new Gson();
        System.out.println(gson.toJson(new SerializedNameValueTest().new User()));

        String str = "{'this is a name':'what????','this is an age':13,'email':'[email protected]'}";
        System.out.println(gson.fromJson(str,User.class));

        String str2 = "{'name':'askldfklaslk','age':-222,'email':'[email protected]'}";
        System.out.println(gson.fromJson(str2, User.class));
    }

    class User
    {
        @SerializedName("this is a name")
        private String name = "123";

        @SerializedName("this is an age")
        private int age = 0;

        private String email = "[email protected]";

        private double[] nums = new double[] { 1.0, 2.0 };

        @Override
        public String toString() {
            return "name:" + name + ",age:" + age + ",eamil:" + email + ",nums:" + Arrays.toString(nums);
        }
    }
}

最后一个反序列化不成功,因为'name'不对应@SerializedName()中的'this is a name'.

b.同时使用value与alternate

同时使用两者可以解决上面的问题,即name字段--->json中的'this is a name'--->name字段:

public class SerializedNameValueAndAlternateTest
{
    public static void main(String[] args) {
        Gson gson = new Gson();
        System.out.println(gson.toJson(new SerializedNameValueAndAlternateTest().new User()));

        String str = "{'this is a name':'what????','this is an age':13,'email':'[email protected]'}";
        System.out.println(gson.fromJson(str, User.class));

        String str2 = "{'name':'askldfklaslk','age':-222,'email':'[email protected]'}";
        System.out.println(gson.fromJson(str2, User.class));
    }

    class User {
        @SerializedName(value = "this is a name",alternate = "name")
        private String name = "123";

        @SerializedName(value = "this is an age",alternate = "age")
        private int age = 0;

        private String email = "[email protected]";

        private double[] nums = new double[] { 1.0, 2.0 };

        @Override
        public String toString() {
            return "name:" + name + ",age:" + age + ",eamil:" + email + ",nums:" + Arrays.toString(nums);
        }
    }
}

alternate就是为反序列化准备的,若找不到value中对应的值,则寻找alternate对应的值,找不到再设为null.

若同时存在value与alternate:

String str3 = "{'name':'altername','this is a name':'value'}";
System.out.println(gson.fromJson(str3, User.class));
String str4 = "{'this is a name':'value','name':'altername'}";
System.out.println(gson.fromJson(str4, User.class));

则以"最晚"出现的值为标准.

(2)@Expose

@Expose可以忽略某个字段,有两个参数:

• serialize
• deserialize

默认情况下都是true,分别表示序列化与反序列化.

System.out.println(gson.toJson(new User()));

class User
{
    @Expose(serialize = false)
    private String name = "123";

    @Expose(deserialize = false)
    private int age = 0;

    @Expose(serialize = false,deserialize = false)
    private String email = "[email protected]";

    @Expose(serialize = true,deserialize = true)
    private double [] nums = new double [] {1.0,2.0};
}

name不允许序列化,age不允许反序列化,email不允许序列化与反序列化,nums既允许也允许反序列化.(此时效果等于没加@Expose)
输出:

咦?好像没用的样子?
因为Gson对象会默认忽略@Expose,想要@Expose生效需要使用Gson Builder:

GsonBuilder gsonBuilder = new GsonBuilder();
gsonBuilder.excludeFieldsWithoutExposeAnnotation();
Gson gson = gsonBuilder.create();
System.out.println(gson.toJson(new User()));

String str = "{'name':'234','age':-3,'email':'23423','nums':[3,4,3]}";
System.out.println(gson.fromJson(str, User.class));

class User
{
    @Expose(serialize = false)
    private String name = "123";

    @Expose(deserialize = false)
    private int age = 0;

    @Expose(serialize = false,deserialize = false)
    private String email = "[email protected]";

    @Expose(serialize = true,deserialize = true)
    private double [] nums = new double [] {1.0,2.0};

    @Override
    public String toString()
    {
        return "name:"+name+",age:"+age+",eamil:"+email+",nums:"+Arrays.toString(nums);
    }
}

不能反序列化age与email,输出的是age与email的默认值.

6.Gson Builder

前面的例子基本上都是通过

Gson gson = new Gson();

来直接实例化一个Gson来使用Gson的,使用Gson Builder可以设置Gson的某些属性,使用其中的create()返回一个Gson.

(1)基础

创建一个简单的Gson,通过create()创建:

GsonBuilder gsonBuilder = new GsonBuilder();
Gson gson = gsonBuilder.create();

(2)命名规则

a.使用系统命名规则

使用

gsonBuilder.setFieldNamingPolicy();

设置json中键的命名规则.6个值可选:

• IDENTITY:相同,json中的键名与字段名相同.
• LOWER_CASE_WITH_DASHES:在原来字段的大写字母前加-,并且把大写变成小写
• LOWER_CASE_WITH_DOTS:在原来字段的大写字母前加.,并且把大写变成小写
• LOWER_CASE_WITHUNDERSCORES:在原来字段的大写字母前加,并且把大写变成小写
• UPPER_CAMEL_CASE:首字母大写
• UPPER_CAMEL_CASE_WITH_SPACES:在原来字段的大写字母前加空格

public class NamingRulesTest
{
    public static void main(String[] args) {
        NamingRulesTest n = new NamingRulesTest();
        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.setFieldNamingPolicy(FieldNamingPolicy.IDENTITY);
        System.out.println(gsonBuilder.create().toJson(n.new User()));

        gsonBuilder.setFieldNamingPolicy(FieldNamingPolicy.LOWER_CASE_WITH_DASHES);
        System.out.println(gsonBuilder.create().toJson(n.new User()));

        gsonBuilder.setFieldNamingPolicy(FieldNamingPolicy.LOWER_CASE_WITH_DOTS);
        System.out.println(gsonBuilder.create().toJson(n.new User()));

        gsonBuilder.setFieldNamingPolicy(FieldNamingPolicy.LOWER_CASE_WITH_UNDERSCORES);
        System.out.println(gsonBuilder.create().toJson(n.new User()));

        gsonBuilder.setFieldNamingPolicy(FieldNamingPolicy.UPPER_CAMEL_CASE);
        System.out.println(gsonBuilder.create().toJson(n.new User()));

        gsonBuilder.setFieldNamingPolicy(FieldNamingPolicy.UPPER_CAMEL_CASE_WITH_SPACES);
        System.out.println(gsonBuilder.create().toJson(n.new User()));
    }

    class User
    {
        @SuppressWarnings("unused")
        private String Name_Name_name = "123";
        @SuppressWarnings("unused")
        private int _age_age_Age_age = 645;
        @SuppressWarnings("unused")
        private double numsNumsNums = 34.45;
    }
}

注意,若某个字段有了@SerializedName,则这个字段遵循@SerializedName的策略.

b.自定义命名规则

重写FieldNamingStrategy中的translateName(Field field),把自定义的FieldNamingStrategy传递给GsonBuilder的setFieldNamingStrategy().
比如想要前缀加上某人的名字:

public class CustomNamingRulesTest
{
    public static void main(String[] args) {
        FieldNamingStrategy myNamingStrategy = new FieldNamingStrategy(){
            @Override
            public String translateName(Field field) {
                return "kr"+field.getName();
            }
        };

        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.setFieldNamingStrategy(myNamingStrategy);
        Gson gson = gsonBuilder.create();
        System.out.println(gson.toJson(new CustomNamingRulesTest().new User()));
    }

    class User
    {
        @SuppressWarnings("unused")
        private String name = "123";
        @SuppressWarnings("unused")
        private int num = 5;
    }
}

(3)null

默认情况下,Gson实例不允许序列化null,如果想要序列化null,借助GsonBuilder的serializeNulls()方法:

public class NullTest
{
    public static void main(String[] args) {
        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.serializeNulls();
        gsonBuilder.setFieldNamingPolicy(FieldNamingPolicy.UPPER_CAMEL_CASE_WITH_SPACES);
        System.out.println(gsonBuilder.create().toJson(new NullTest().new User()));
    }

    class User
    {
        @SuppressWarnings("unused")
        private String thisIsANullField;
        @SuppressWarnings("unused")
        private Integer andThisIsANullFieldToo;
    }
}

(4)排除

前面已经接触到了@Expose这样的排除策略,下面看看更加强大的配合GsonBuilder使用的排除策略.主要有四种:属性名排除,类型排除,修饰符排除,@Expose排除.

a.属性名排除

public class ExclusionNameTest
{
    public static void main(String[] args) {
        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.serializeNulls();
        ExclusionStrategy exclusionName = new ExclusionStrategy(){
            @Override
            public boolean shouldSkipField(FieldAttributes f)
            {
                return f.getName().endsWith("e");
            }

            @Override
            public boolean shouldSkipClass(Class<?> arg0) {
                return false;
            }
        };
        gsonBuilder.setExclusionStrategies(exclusionName);

        System.out.println(gsonBuilder.create().toJson(new ExclusionNameTest().new User()));
    }

    class User
    {
        private String name;
        private int num;
    }
}

排除字段名以e结尾的字段.下面是重点:

public boolean shouldSkipField(FieldAttributes f)
{
    return f.getName().endsWith("e");
}

重写的shouldSkipField从名字可以看出跳过某些字段,返回true表示跳过,即排除这个字段.上面的例子中若名字以e结尾则跳过,因此输出:

b.类型排除

public class ExclusionTypeTest
{
    public static void main(String[] args) {
        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.serializeNulls();
        ExclusionStrategy exclusionType = new ExclusionStrategy(){

            @Override
            public boolean shouldSkipField(FieldAttributes arg0) {
                return false;
            }

            @Override
            public boolean shouldSkipClass(Class<?> cls) {
                return cls == String.class;
            }
        };
        gsonBuilder.setExclusionStrategies(exclusionType);

        System.out.println(gsonBuilder.create().toJson(new ExclusionNameTest().new User()));
    }

    class User
    {
        @SuppressWarnings("unused")
        private String name;
        @SuppressWarnings("unused")
        private int num;
    }
}

重写的shouldSkipClass表示要跳过的类,这里跳过了String,只剩下num.

@Override
public boolean shouldSkipClass(Class<?> cls) {
    return cls == String.class;
}

c.修饰符排除

这个不用重写方法了,直接使用GsonBuilder的excludeFieldsWithModifiers(),参数是java.lang.reflect.Modifier:

这个可选的比较多就不一一列举了,只选了一个final的例子:

public class ExclusionModifierTest
{
    public static void main(String[] args) {
        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.excludeFieldsWithModifiers(Modifier.FINAL);
        System.out.println(gsonBuilder.create().toJson(new User()));
    }
}

class User
{
    @SuppressWarnings("unused")
    private final String name = "123";
    @SuppressWarnings("unused")
    private static int num;
}

d.@Expose排除

这个准确来说是排除没有被@Expose注解的字段:

public class ExclusionExposeTest
{
    public static void main(String[] args) {
        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.excludeFieldsWithoutExposeAnnotation();
        System.out.println(gsonBuilder.create().toJson(new ExclusionExposeTest().new User()));
    }

    class User
    {
        @SuppressWarnings("unused")
        private String name;
        @SuppressWarnings("unused")
        private int num;
        @Expose(serialize = true,deserialize = true)
        private int age;
    }
}

(5)Lenient

这是有关于反序列化时JSON的容错机制的问题,因为通常来说将一个对象通过Gson转换成json没什么大问题,但是将json转换为对象的时候就...就难说了,因为不知道是否符合标准的json格式,因此Gson提供了一定的容错机制,就是Lenient.

Lenient翻译过来是"宽容的"的意思,可以通过:

GsonBuilder gsonBuilder = new GsonBuilder();
gsonBuilder.setLenient();

开启.
Gson内部有一个JsonReader类,默认情况下,JsonReader是严格的且仅接受符合RFC 4627标准的json(RFC 4627标准可以看看这里),设置为lenient后可以"容忍"以下几种错误:

• a.以 ) ] } ' \n 开头
• b.多个顶层值.
• c.任何类型的顶层值
• d.数字可能是NaNs或infinites
• e.以行注释//或#结尾
• f.以一个换行符结束
• g.C风格的注释/**/,可能会嵌套
• h.键/字符串没有引号或者单引号
• i.数组元素以;分隔
• j.不必要的数组分隔符,"默认"null为省略值,比如[1,,2],"默认"第二个元素为null
• k.键值以=或=>分隔而非使用:
• l.键值对以;分隔而非使用,

设置setLenient()后,Gson会尽可能解析有错误的json,若实在无能为力,会抛出MalformedJsonException异常.
下面是一个不严格的json的例子:

public class LenientTest
{
    public static void main(String[] args)
    {
        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.setLenient();

        String str = "{'name'='234';'num'=6}";
        System.out.println(gsonBuilder.create().fromJson(str,User.class));
    }

    class User
    {
        private String name;
        private int num;

        @Override
        public String toString()
        {
            return "name:"+name+",num:"+num;
        }
    }
}

下面是一个异常的例子:

String errorStr = "{'name'=????,,,,,,}";
System.out.println(gsonBuilder.create().fromJson(errorStr,User.class));

(6)Floats & Doubles

先看一个例子:

public class FloatTest
{
    public static void main(String[] args) {
        Gson gson = new Gson();
        System.out.println(gson.toJson(new FloatTest().new User()));
    }

    class User
    {
        @SuppressWarnings("unused")
        private float f = Float.POSITIVE_INFINITY;
    }
}

会报错:

它说Infinity在JSON标准是一个非法值.想要重写这个行为需要使用GsonBuilder.serializeSpecialFloatingPointValues().
因为JSON规范不允许NaN,-Infinity,Infinity,因此会报错.下面使用GsonBuilder:

public class FloatTest
{
    public static void main(String[] args) {
        // Gson gson = new Gson();
        // System.out.println(gson.toJson(new FloatTest().new User()));

        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.serializeSpecialFloatingPointValues();
        System.out.println(gsonBuilder.create().toJson(new FloatTest().new User()));
    }

    class User
    {
        @SuppressWarnings("unused")
        private float f = Float.POSITIVE_INFINITY;
    }
}

直接输出Infinity:

double也类似:

public class DoubleTest
{
    public static void main(String[] args)
    {
        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.serializeSpecialFloatingPointValues();
        System.out.println(gsonBuilder.create().toJson(new DoubleTest().new User()));
    }

    class User
    {
        @SuppressWarnings("unused")
        private double ni = Double.NEGATIVE_INFINITY;
        @SuppressWarnings("unused")
        private double pi = Double.POSITIVE_INFINITY;
        @SuppressWarnings("unused")
        private double nan = Double.NaN;
    }
}

(7)模型版本化

通过@Since与@Until添加版本控制,控制某个版本在序列化与反序列化时忽略或忽略某个字段.
@Since表示从某个版本开始这个字段不忽略,@Until表示这个版本后将忽略该字段.

需要配合GsonBuilder的setVersion使用,设定版本号.

public class VersionTest
{
    public static void main(String[] args) {
        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.serializeNulls();
        System.out.println(gsonBuilder.setVersion(0.9).create().toJson(new VersionTest().new User()));
        System.out.println(gsonBuilder.setVersion(1.0).create().toJson(new VersionTest().new User()));
        System.out.println(gsonBuilder.setVersion(1.4).create().toJson(new VersionTest().new User()));
        System.out.println(gsonBuilder.setVersion(1.5).create().toJson(new VersionTest().new User()));
    }

    class User
    {
        @Since(1.0)
        private String name;
        @Until(1.4)
        private int num;
    }
}

@Since的范围包含了左区间端点,@Until的范围不包含右区间端点.

(8)格式化日期

可以使用setDateForamt()来格式化日期输出:

gsonBuilder.setDateFormat(int style);
gsonBuilder.setDateFormat(String pattern);
gsonBuilder.setDateFormat(int dateStyle,int timeStyle);

第一个函数的参数为DateFormat中的常量值,第二个函数的参数是类似SimpleDateFormat中的String,第三个函数的参数与第一个类似,分开日期与时间设置.
再说以下setPrettyPrinting()这个函数,看名字就知道,美化打印的.会加上空格.

public class FormatTest
{
    public static void main(String[] args) {
        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.setPrettyPrinting();

        System.out.println(gsonBuilder.setDateFormat(DateFormat.SHORT, DateFormat.SHORT).create().toJson(new FormatTest().new User()));
        System.out.println(gsonBuilder.setDateFormat(DateFormat.SHORT,DateFormat.LONG).create().toJson(new FormatTest().new User()));
        System.out.println(gsonBuilder.setDateFormat(DateFormat.MEDIUM, DateFormat.MEDIUM).create().toJson(new FormatTest().new User()));
        System.out.println(gsonBuilder.setDateFormat(DateFormat.LONG, DateFormat.MEDIUM).create().toJson(new FormatTest().new User()));
        System.out.println(gsonBuilder.setDateFormat(DateFormat.FULL, DateFormat.FULL).create().toJson(new FormatTest().new User()));
        System.out.println(gsonBuilder.setDateFormat("yyyy-MM-dd HH:mm:ss").create().toJson(new FormatTest().new User()));
    }

    class User
    {
        @SuppressWarnings("unused")
        private Date date = new Date();
    }
}

7.Gson进阶

(1)枚举

枚举其实也与普通的字段类似,序列化与反序列化正常操作.

public class EnumTest
{
    public static void main(String[] args) {
        Gson gson = new Gson();
        System.out.println(gson.toJson(new EnumTest().new User()));

        String str = "{\"season\":\"SUMMER\",\"name\":\"8888\"}";
        System.out.println(gson.fromJson(str,User.class));
    }

    enum Season
    {
        SPRING,
        SUMMER,
        AUTUMN,
        WINTER
    }

    class User
    {
        private Season season = Season.SPRING;
        private String name = "123";

        @Override
        public String toString()
        {
            return "season:"+season.toString()+",name:"+name;
        }
    }
}

主要配合@SerializedName()使用,例如对于月份,可以使用1,2,3来表示:

public class EnumTest
{
    public static void main(String[] args) {
        Gson gson = new Gson();
        System.out.println(gson.toJson(new EnumTest().new User()));

        String str = "{\"season\":\"SUMMER\",\"name\":\"8888\",\"month\":\"2\"}";
        System.out.println(gson.fromJson(str,User.class));
    }

    enum Season
    {
        SPRING,
        SUMMER,
        AUTUMN,
        WINTER
    }

    class User
    {
        private Season season = Season.SPRING;
        private String name = "123";
        private Month month = Month.January;

        @Override
        public String toString()
        {
            return "season:"+season.toString()+",name:"+name+",month:"+month.toString();
        }
    }

    enum Month
    {
        @SerializedName("1")
        January,
        @SerializedName("2")
        February,
        @SerializedName("3")
        March,
        @SerializedName("4")
        April,
        @SerializedName("5")
        May,
        @SerializedName("6")
        June,
        @SerializedName("7")
        July,
        @SerializedName("8")
        August,
        @SerializedName("9")
        September,
        @SerializedName("10")
        October,
        @SerializedName("11")
        November,
        @SerializedName("12")
        December
    }
}

(2)泛型

这里主要说一下反序列化泛型.因为序列化泛型...直接toJson()就好了.
想想这样的情景:有一个待解析的Json String,类型为List<T>,因此想有一个这样的方法:

public <T> List<T> fromJSON(String json,Class<T> cls);

好了,怎么写呢,之前用的是TypeToken()实现反序列化:

List<Integer> integerList = new ArrayList<>();
integerList.add(3);
integerList.add(-999);
gson.toJson(integerList, new TypeToken<List<Integer>>() {}.getType());

答案就是使用ParameterizedType:重写里面的getActualTypeArguments(),getOwnerTpye()与getRawType():

public class GenericsTest
{
    public static void main(String[] args) {
        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.setPrettyPrinting();
        Gson gson = gsonBuilder.create();

        List<Integer> integerList = new ArrayList<>();
        integerList.add(3);
        integerList.add(-999);
        gson.toJson(integerList, new TypeToken<List<Integer>>() {}.getType());

        List<String> stringList = new ArrayList<>();
        stringList.add("12321");
        stringList.add("(@)#$@#");
        String t1 = gson.toJson(integerList);
        String t2 = gson.toJson(stringList);

        new GenericsTest().fromJSON(t1, Integer.class).stream().forEach(System.out::println);
        new GenericsTest().fromJSON(t2, String.class).stream().forEach(System.out::println);
    }

    public <T> List<T> fromJSON(String json,Class<T> cls)
    {
        return new Gson().fromJson(json,new ParameterizedTypeImpl(cls));
    }

    class ParameterizedTypeImpl implements ParameterizedType
    {
        private Class cls;
        public ParameterizedTypeImpl(Class cls)
        {
            this.cls = cls;
        }

        @Override
        public Type[] getActualTypeArguments()
        {
            return new Type[]{cls};
        }

        @Override
        public Type getRawType()
        {
            return List.class;
        }

        @Override
        public Type getOwnerType()
        {
            return null;
        }
    }
}

@Override
public Type[] getActualTypeArguments()
{
    return new Type[]{cls};
}

返回实际参数类型数组,在这里是String.class与Integer.class.

@Override
public Type getRawType()
{
    return List.class;
}

返回声明这个类型的类或接口,在这里是List.class.

@Override
public Type getOwnerType()
{
    return null;
}

返回其成员之一的类型,就是说如果完全限定类名为A.B,则返回A,在这里没有A,因此为null.
定义好实现ParameterizedTpye接口的类后,把它传递给fromJson()作为第二参数,构造方法的参数为List<T>中的T.class.

(4)自定义序列化

自定义序列化一般用于自定义简化json.
比如有一个User类:

class User
{
    private String name = "123";
    private String email = "[email protected]";
    private int [] nums = new int [] {1,2,3};
}

想要不序列化name,可以使用@Expose(serialize = false),但是想要部分序列化nums,比如只是需要第一个nums[0],上面的@Expose,@SerializedName等注解都用不上,这时需要使用JsonSerializer自定义序列化,重写其中的:

public JsonElement serialize(T t,Type type,JsonSerializationContext context)

可以返回一个JsonObject,这里的JsonObject可以自定义添加属性(即键值对).
需要配合GsonBuilder使用,创建了自己的JsonSerializer<T>后,把它传递给GsonBuilder的registerTypeAdapter():

public class SerializeCustomTest
{
    public static void main(String[] args) {
        GsonBuilder gsonBuilder = new GsonBuilder();
        JsonSerializer<User> serializer = new JsonSerializer<SerializeCustomTest.User>() {
            @Override
            public JsonElement serialize(User u,Type type,JsonSerializationContext context)
            {
                JsonObject object = new JsonObject();
                object.addProperty("name",u.getName());
                object.addProperty("email",u.getEmail());
                object.addProperty("nums",u.getNums()[0]);
                return object;
            }
        };
        gsonBuilder.registerTypeAdapter(User.class,serializer);
        System.out.println(gsonBuilder.create().toJson(new SerializeCustomTest().new User()));
    }

    class User
    {
        private String name = "123";
        private String email = "[email protected]";
        private int [] nums = new int [] {1,2,3};

        public int [] getNums()
        {
            return nums;
        }

        public String getName()
        {
            return name;
        }

        public String getEmail()
        {
            return email;
        }
    }
}

(5)自定义反序列化

自定义反序列化主要就是针对有多余键值对的json,比如User只需要name字段,但是传来的json包含了诸如age这样的键,因此只需挑选所需的进行反序列化.与序列化类似,首先定义自己的JsonDeserializer<T>,重写其中的:

public T deserialize(JsonElement,Type,JsonDeserializationContext);

返回一个自定义的反序列化的对象.最后在GsonBuilder中registerTypeAdapter()即可.

public class DeserializedCustomTest
{
    public static void main(String[] args) {
        GsonBuilder gsonBuilder = new GsonBuilder();
        JsonDeserializer<User> deserializer = new JsonDeserializer<DeserializedCustomTest.User>() {
            @Override
            public User deserialize(JsonElement json,Type type,JsonDeserializationContext context)
            {
                JsonObject object = json.getAsJsonObject();
                User user = new DeserializedCustomTest().new User(object.get("name").getAsString());
                return user;
            }
        };
        gsonBuilder.registerTypeAdapter(User.class, deserializer);

        String str = "{\"aaa\":\"bbbb\",\"name\":\"this is a name\",\"age\":\"444\"}";
        System.out.println(gsonBuilder.create().fromJson(str,User.class));
    }

    class User
    {
        private String name;
        private int age;

        public User(String name)
        {
            this.name = name;
        }

        @Override
        public String toString()
        {
            return "name:"+name+",age:"+age;
        }
    }
}

(6)默认实例

可以设置反序列化时的默认值,通过构造方法实现.比如json中没有所需要的字段的值,默认情况下为null,如果想要不为null,可以设定默认值,对于对象可以设定构造方法.

通过实现InstanceCreator<T>来实现,重写其中的createInstance方法,再配合GsonBuilder的registerTypeAdapter().

public class InstanceCustomTest
{
    public static void main(String[] args) {
        GsonBuilder gsonBuilder = new GsonBuilder();
        gsonBuilder.registerTypeAdapter(
            AA.class,
            new InstanceCustomTest().new AAInstance(
                new InstanceCustomTest().new BB("123","456")
            )
        );

        String str = "{\"name\":\"123123\"}";
        System.out.println(gsonBuilder.create().fromJson(str, AA.class));
    }

    class AAInstance implements InstanceCreator<AA>
    {
        private BB bb;
        public AAInstance(BB bb)
        {
            this.bb = bb;
        }

        @Override
        public AA createInstance(Type type)
        {
            return new AA(bb);
        }
    }

    class AA
    {
        private String name = "AA";
        private BB bb;

        public AA(BB bb)
        {
            this.bb = bb;
        }

        @Override
        public String toString()
        {
            return "name:"+name+",bb:"+bb.toString();
        }
    }

    class BB
    {
        private String field1;
        private String field2;

        public BB(String field1,String field2)
        {
            this.field1 = field1;
            this.field2 = field2;
        }

        @Override
        public String toString()
        {
            return "field1:"+field1+",field2:"+field2;
        }
    }
}

json没有为BB类设定值,采用默认值.

(7)@JsonAdapter

这个是自定义序列化/反序列化的注解,可以简化JsonSerializer与JsonDeserializer.另外不用配合GsonBuilder使用,直接使用new Gson().toJson()/fromJson()即可,比直接使用JsonSerializer与JsonDeserializer要简单.
首先使一个类实现JsonSerializer<T>接口,接着把这个类作为要注解的字段的@JsonAdapter的参数:

class Serializer implements JsonSerializer<User>
class ContainUser
{
    private String userId = "x";
    @JsonAdapter(Serializer.class)
    private User user = new User();
}

public class JsonAdapterSerializeTest
{
    public static void main(String[] args)
    {
        Gson gson = new Gson();
        System.out.println(gson.toJson(new JsonAdapterSerializeTest().new ContainUser()));
    }

    class Serializer implements JsonSerializer<User>
    {
        @Override
        public JsonElement serialize(User u,Type type,JsonSerializationContext context)
        {
            JsonObject object = new JsonObject();
            object.addProperty("name",u.getName());
            object.addProperty("email",u.getEmail());
            object.addProperty("nums",u.getNums()[0]);
            return object;
        }
    }

    class ContainUser
    {
        private String userId = "x";
        @JsonAdapter(Serializer.class)
        private User user = new User();
    }

    class User
    {
        private String name = "123";
        private String email = "[email protected]";
        private int [] nums = new int [] {1,2,3};

        public int [] getNums()
        {
            return nums;
        }

        public String getName()
        {
            return name;
        }

        public String getEmail()
        {
            return email;
        }
    }
}

反序列化也类似.

public class JsonAdapterDeserializeTest
{
    public static void main(String[] args) {
        Gson gson = new Gson();
        String str = "{\"aaaa\":\"bbbb\",\"name\":\"this is a name\",\"age\":\"444\"}";
        System.out.println(gson.fromJson(str,User.class));
    }

    class Deserialize implements JsonDeserializer<User>
    {
        @Override
        public User deserialize(JsonElement json,Type type,JsonDeserializationContext context)
        {
            JsonObject object = json.getAsJsonObject();
            User user = new JsonAdapterDeserializeTest().new User(object.get("name").getAsString());
            return user;
        }
    }

    @JsonAdapter(Deserialize.class)
    class User
    {
        private String name = "123";
        public User()
        {

        }
        public User(String name)
        {
            this.name = name;
        }

        @Override
        public String toString()
        {
            return "name:"+name;
        }
    }
}

实现JsonDeserializer<T>,重写deserialize(),在所需的类上添加注解.

注意@JsonAdapter注解只能添加在类上.

8.源码

所有例子的完整源码.
github.
码云.
Gson官方github.

9.参考链接

1.Gson Tutorial Series by Future Studio

2.gson完全教程

3.gson教程

4.gson用户指南

5.Gson User guide

6.gson