spray json in scala: 反序列化 json 未知字段而不丢失它们

spray json in scala: deserializing json with unknown fields without losing them

我正在寻找 的解决方案,但对于 json spray 和我的搜索以及尝试使用 json spray 到目前为止都失败了。

如果我有以下 json:

{
  "personalData": {
    "person": {
        "first": "first_name",
        "last": "last_name"
    },
    "contact": {
        "phone": "1111111",
        "address": {
            "line": "123 Main St",
            "city": "New York"
        }
    },
    "xx": "yy", // unknown in advanced
    "zz": { // unknown in advanced
        "aa": "aa",
        "bb": "bb",
        "cc": {
            "dd": "dd",
            "ee": "ee"
        }
    }
  }
}

我们确定 json 将包含人员和联系人,但我们不知道在我们上游可能会创建哪些我们不关心的其他字段 about/use。

我想将此 JSON 序列化为一个包含人员和联系人的案例 class,但另一方面,我不想丢失其他字段(将它们保存在地图中,以便class 将被反序列化为与接收到的相同的 json。

这是我完成的程度:

case class Address(
                    line:        String,
                    city:        String,
                    postalCode:  Option[String]
                  )

case class Contact(
                    phone:       String,
                    address:     Address
                  )

case class Person(
                   first:      String,
                   last:      String
                 )

case class PersonalData(
                         person:      Person,
                         contact:     Contact,
                         extra:       Map[String, JsValue]
                       )

implicit val personFormat = jsonFormat2(Person)
implicit val addressFormat = jsonFormat3(Address)
implicit val contactFormat = jsonFormat2(Contact)

implicit val personalDataFormat = new RootJsonFormat[PersonalData] {
  def write(personalData: PersonalData): JsValue = {
    JsObject(
      "person" -> personalData.person.toJson,
      "contact" -> personalData.contact.toJson,
      // NOT SURE HOW TO REPRESENT extra input
    )
  }

  def read(value: JsValue): CAERequestBEP = ???
}

谁能帮我用 spray.json 而不是玩游戏?我花了很长时间尝试这样做,但似乎无法让它发挥作用。

为此,您需要为 PersonalDataFormat 编写自己的格式化程序:

case class Person(first: String, last: String)
case class Address(line: String, city: String)
case class Contact(phone: String, address: Address)
case class PersonalData(person: Person, contact: Contact, extra: Map[String, JsValue])
case class Entity(personalData: PersonalData)

implicit val personFormat = jsonFormat2(Person)
implicit val addressFormat = jsonFormat2(Address)
implicit val contactFormat = jsonFormat2(Contact)
implicit object PersonalDataFormat extends RootJsonFormat[PersonalData] {
  override def read(json: JsValue): PersonalData = {
    val fields = json.asJsObject.fields
    val person = fields.get("person").map(_.convertTo[Person]).getOrElse(???) // Do error handling instead of ???
    val contact = fields.get("contact").map(_.convertTo[Contact]).getOrElse(???) // Do error handling instead of ???
    PersonalData(person, contact, fields - "person" - "contact")
  }

  override def write(personalData: PersonalData): JsValue = {
    JsObject(personalData.extra ++ ("person" -> personalData.person.toJson, "contact" -> personalData.contact.toJson))
  }
}

implicit val entityFormat = jsonFormat1(Entity)

val jsonResult = jsonString.parseJson.convertTo[Entity]

结果是:

Entity(PersonalData(Person(first_name,last_name),Contact(1111111,Address(123 Main St,New York)),Map(xx -> "yy", zz -> {"aa":"aa","bb":"bb","cc":{}})))

(假设 json 不完全是上面的 json,而是一个有效的相似的)

代码 运行 Scastie