預備知識

json4s的數據結構AST (Abstract Syntax Tree)。

sealed abstract class JValue
case object JNothing extends JValue // 'zero' for JValue
case object JNull extends JValue
case class JString(s: String) extends JValue
case class JDouble(num: Double) extends JValue
case class JDecimal(num: BigDecimal) extends JValue
case class JInt(num: BigInt) extends JValue
case class JBool(value: Boolean) extends JValue
case class JObject(obj: List[JField]) extends JValue
case class JArray(arr: List[JValue]) extends JValuetype JField = (String, JValue)

我們可以通過 json4s 對json所做的操作如下圖所示，中間為 Json AST (簡單理解就是一個用JValue表示的 JSON)。
在這里插入圖片描述

另外，org.json4s下定義了很多scala原生數據轉JValue的隱式轉換（即多數操作下可以把原生數據當做JValue直接使用）
在這里插入圖片描述
但是注意：Tuple不能自動轉為JValue，在需要轉換的時候，render先轉一下，如：json merge render("height",175)

一、創建json對象

{"name":"luca", "id": "1q2w3e4r5t", "age": 26, "url":"http://www.nosqlnocry.wordpress.com"}

方式1：parse函數

// parse from string example，得到的是JValue對象
var json = parse("""{"name":"luca", "id": "1q2w3e4r5t", "age": 26, "url":"http://www.nosqlnocry.wordpress.com"}""")

方式2：dsl創建

org.json4s中定義了從tuple到JValue的操作符

// DSL example，是JObject對象
var json = ("name","luca") ~ ("id","1q2w3e4r5t") ~ ("age",26) ~ ("url","http://www.nosqlnocry.wordpress.com")
// tuples can be defined also like this: ("id" -> "1q2w3e4r5t")

println(json)
JObject(List((name,JString(luca)), (id,JString(1q2w3e4r5t)), (age,JInt(26)), (url,JString(http://www.nosqlnocry.wordpress.com))))

二、常用操作

2.1 新增一個field

注意，以下兩個方法都不會覆蓋，若原json中已有height，則新json中會有兩個height.

//法1：使用dsl，JObject才有~方法
json = json ~ ("height" -> 175) 
//法2: 使用merge,要求類型和左邊一致，所以json為parse出來的JValue時，要用render生成JValue再merge
json = json merge render("height",175)

2.2 更新一個field

使用 transformField

json = json transformField {case JField("name", _) => ("NAME", JString("Luca")) //還可重命名case JField("age", JInt(age)) => ("age", JInt(age+1))//更新值
}json = json merge render("age",26) //若json 中已有"age"

2.3 刪除一個field

json = json removeField {case JField("NAME", _) => true //被刪除case _ => false
}
// 或等價的
json = json filterField {case JField("NAME", _) => false case _ => true //被保留}

2.4 獲取一個field

println(compact(json \\ "age")) // 27 嵌套獲取-見下
println(compact(json \ "age"))  // 27
println(compact(json.children(1))) // 27

三、高階操作

{"name": "luca","id": "1q2w3e4r5t","age": 26,"url": "http://www.nosqlnocry.wordpress.com","url": "https://nosqlnocry.wordpress.com","loginTimeStamps": [1434904257,1400689856,1396629056],"messages": [{"id": 1,"content": "Please like this post!"},{"id": 2,"content": "Forza Roma!"}],"profile": {"id": "my-nickname","score": 123,"avatar": "path.jpg"}
}

3.1 選取field

println(JSON)
//JObject(List((name,JString(luca)), (id,JString(1q2w3e4r5t)), (age,JInt(26)), (url,JString(http://www.nosqlnocry.wordpress.com)), (url,JString(https://nosqlnocry.wordpress.com)), (loginTimeStamps,JArray(List(JInt(1434904257), JInt(1400689856), JInt(1396629056)))), (messages,JArray(List(JObject(List((id,JInt(1)), (content,JString(Please like this post!)))), JObject(List((id,JInt(2)), (content,JString(Forza Roma!))))))), (profile,JObject(List((id,JString(my-nickname)), (score,JInt(123)), (avatar,JString(path.jpg)))))))println(JSON\\"id") //獲取所有嵌套的id數據
// prints: JObject(List((id,JString(1q2w3e4r5t)), (id,JInt(1)), (id,JInt(2)), ...println(JSON\"id")//獲取第一層的id數據
// prints: JString(1q2w3e4r5t)println(JSON\"url") //如果第一層有多個，則返回JArray
// prints: JArray(List(JString(http://www...), JString(https://nosqlnocry...val messagesIds = (JSON \ "messages") \ "id" //獲取JAray中的id數據
println(messagesIds)
// prints: JArray(List(JInt(1), JInt(2)))
println(messagesIds.values)
// prints: List(1,2)
//或用for語句
val messagesIds2=  for {JObject(child) <- JSONJField("id", JInt(id)) <- child} yield idprintln(messagesIds2)// prints: List(1,2)

for語句的<-在JValue中做了特殊處理，會返回所有匹配項。

    for {JObject(child) <- JSON //這回匹配所有JObject，不管是不是嵌套}{println(child)}
//List((name,JString(luca)), (id,JString(1q2w3e4r5t)), (age,JInt(26)), (url,JString(http://www.nosqlnocry.wordpress.com)), (url,JString(https://nosqlnocry.wordpress.com)), (loginTimeStamps,JArray(List(JInt(1434904257), JInt(1400689856), JInt(1396629056)))), (messages,JArray(List(JObject(List((id,JInt(1)), (content,JString(Please like this post!)))), JObject(List((id,JInt(2)), (content,JString(Forza Roma!))))))), (profile,JObject(List((id,JString(my-nickname)), (score,JInt(123)), (avatar,JString(path.jpg))))))
//List((id,JInt(1)), (content,JString(Please like this post!)))
//List((id,JInt(2)), (content,JString(Forza Roma!)))
//List((id,JString(my-nickname)), (score,JInt(123)), (avatar,JString(path.jpg)))

JValue的<-調用的是

    def foreach(f: JValue => Unit): Unit =self.filter(p).foreach(f)

Array的<-調用的是

  def foreach[U](f: A => U): Unit = {var i = 0val len = lengthwhile (i < len) { f(this(i)); i += 1 }}

3.2 取出field

println(compact(render(JSON \ "messages")))
// prints: [{"id":1,"content":"Please like this post!"},{"id":2,"content":"Forza Roma!"}]
println(pretty(render((JSON \ "messages")\"content")))
// prints: [ "Please like this post!", "Forza Roma!" ] // note it is not compacted anymoreprintln(pretty(render(JSON \ "age")))
// prints: 26println(compact(render(JSON \ "name")))
// prints: "luca" // note the apostrophesvar name = for { JString(x) <- (JSON \\ "name") } yield x //或用for表達式去掉雙引號
println(name(0))
// prints: lucavar name = (JSON \ "name") //或用values去掉雙引號，保留對應基本類型時，推薦這種方法
println(name.values)
// prints: lucaimplicit val formats = DefaultFormats
val name = (JSON \ "name").extract[String]//或直接extract，已知需要的類型時推薦這種方法
println(name)
// prints: luca

name.values 原理：在這里插入圖片描述

3.3 查找和過濾filed

//返回第一個遇到的元素
val URL = JSON findField {
case JField("url", _) => true
case _ => false
}
println(URL)
// prints: Some((url,JString(http://www.nosqlnocry.wordpress.com)))// 返回所有符合條件的元素
val URLs = JSON filterField {
case JField("url", _) => true
case _ => false
}
println(URLs)
// prints: List((url,JString(http://www.nosqlnocry...)), (url,JString(https://nosqlnocry...)

3.4 合并與差異另一個Json2：merge和diff

{"messages": [{"id": 3,"content": "how to merge?"}],"url": "anotherURL","loginTimeStamps": 1400689856,"profile": {"avatar": "new.jpg"},"new": "new value"
}

Json1 merge Json2

如果字段法Json1/f1與Json2/f1結構不同，或者僅具有簡單結構，則Json2會替換Json1的f1
若結構相同且為復雜結構，則會合并
若Json2/f1在Json1中不存在，則新增

diff 獲取兩個JSon間的不同（用得少）：

val newUserJSON = """{"name":"luca","id": "anotherID","age": 26,"url":"http://www.nosqlnocry.wordpress.com",               "profile":{"id":"another-nickname", "score":99999, "avatar":"path.jpg"}
}
"""    
val Diff(changed, added, deleted) = JSON diff parse(newUserJSON)println(compact(render(changed)))
println(added)
println(pretty(render(deleted)))
/* print:
{"id":"anotherID","profile":{"id":"another-nickname","score":99999}}
JNothing
{"url" : "https://nosqlnocry.wordpress.com","loginTimeStamps" : [ 1434904257, 1400689856, 1396629056 ],"messages" : [ {"id" : 1,"content" : "Please like this post!"}, {"id" : 2,"content" : "Forza Roma!"} ]
}*/

3.5 類和JSon間的轉換：decompose, extract, write和read

case class Item(info: String, rank: Int)
case class Item2(info: String, rank: Int, name:String)
implicit val formats: Formats = DefaultFormatsval vMap=Map("info" -> "abcd", "rank" -> 123, "other" -> "dsf")
val jsonStr = write(vMap)
println(jsonStr)
//{"info":"abcd","rank":123,"other":"dsf"}val json = parse(jsonStr)
println(json)val json2 = Extraction.decompose(vMap)//可以理解為等價于parse(write(vMap)）
println(json2)val json=parse(jsonStr)
//val json2=
println(json.extract[Map[String,Any]])
//Map(info -> abcd, rank -> 123, other -> dsf)
println(read[Map[String,Any]](jsonStr))//可理解為和json.extract效果一樣，但是跳過了將str轉為JValue對象的過程
//Map(info -> abcd, rank -> 123, other -> dsf)println(json.extract[Item])//case class 的字段名要和json的field一致，可少不可多與json有的field
//Item(abcd,123)
println(read[Item](jsonStr))
//Item(abcd,123)println(json.extract[Item2])//不可多于json有的field
//報錯，org.json4s.MappingException: No usable value for nameprintln(read[Item2](jsonStr))
//報錯，org.json4s.MappingException: No usable value for name

不用默認格式：(非scala基類作為父類的話，默認格式解析會出錯）

  trait Animalcase class Dog(name: String) extends Animalcase class Fish(weight: Double) extends Animalcase class Animals(animals: List[Animal])implicit val formats1: Formats = DefaultFormatsval formats2: Formats = Serialization.formats(ShortTypeHints(List(classOf[Dog], classOf[Fish])))implicit val mf = manifest[Animals]val ser1 = write(Animals(Dog("pluto") :: Fish(1.2) :: Nil))(formats1)val ser2 = write(Animals(Dog("pluto") :: Fish(1.2) :: Nil))(formats2)println(ser1)//{"animals":[{"name":"pluto"},{"weight":1.2}]}println(ser2)//{"animals":[{"jsonClass":"BasicTest$Dog","name":"pluto"},{"jsonClass":"BasicTest$Fish","weight":1.2}]}println(read[Animals](ser2)(formats2, mf))//Animals(List(Dog(pluto), Fish(1.2)))println(parse(ser2).extract[Animals](formats2,mf))//Animals(List(Dog(pluto), Fish(1.2)))println( read[Animals](ser2)(formats1,mf))// 報錯//org.json4s.MappingException: No usable value for animals，No constructor for type Animal, JObject(List((jsonClass,JString(BasicTest$Dog)), (name,JString(pluto))))println( read[Animals](ser1))//等價于println( read[Animals](ser1)(formats1,mf)) ，報錯//org.json4s.MappingException: No usable value for animals No constructor for type Animal, JObject(List((name,JString(pluto))))println(parse(ser1).extract[Animals])//報錯//org.json4s.MappingException: No usable value for animals No constructor for type Animal, JObject(List((name,JString(pluto))))println(parse(ser2).extract[Animals])//報錯//org.json4s.MappingException: No constructor for type Animal, JObject(List((jsonClass,JString(BasicTest$Dog)), (name,JString(pluto))))

參考

官方教程
WORKING WITH JSON IN SCALA USING THE JSON4S LIBRARY (PART ONE)
WORKING WITH JSON IN SCALA USING THE JSON4S LIBRARY (PART TWO)
Purpose of render in json4s