hello.scala

  case class Field(field: String, value: Any)
  case class Fields(fields: List[Field], score: Double, siteId: String, id: String)

  def hello: Unit = {
    val rdd = sc.textFile("hello.json").map { json =>
      import org.json4s.jackson.JsonMethods._
      implicit val jsonFormats = DefaultFormats
      val fields               = parse(json).extract[Fields]
      for {
        field <- fields.fields if field.field == "body"
        body <- field.value match {
          case a: List[String] => a
          case v: String => List(v)
        }
        word <- body.split(" ")
      } yield {
        word
      }
    }

    val collected = rdd.collect()
    println("Words:" + collected)
    collected
  }