Scala 2.11 + Play Framework 2.3 Case 类和函数中的 22 个字段限制

Question

Scala 2.11 已发布，案例类的 22 个字段限制似乎已修复（ Scala 问题，发行说明）。

这对我来说已经有一段时间了，因为我使用案例类来建模在 Play + Postgres Async中有超过 22 个字段的数据库实体。 我在 Scala 2.10 中的解决方案是将模型分解为多个案例类，但我发现这个解决方案难以维护和扩展，我希望在切换到 Play 2.3.0-RC1 + Scala 2.11 后我可以实现如下所述的内容。 0：

package entities

case class MyDbEntity(
  id: String,
  field1: String,
  field2: Boolean,
  field3: String,
  field4: String,
  field5: String,
  field6: String,
  field7: String,
  field8: String,
  field9: String,
  field10: String,
  field11: String,
  field12: String,
  field13: String,
  field14: String,
  field15: String,
  field16: String,
  field17: String,
  field18: String,
  field19: String,
  field20: String,
  field21: String,
  field22: String,
  field23: String,
) 

object MyDbEntity {
  import play.api.libs.json.Json
  import play.api.data._
  import play.api.data.Forms._

  implicit val entityReads = Json.reads[MyDbEntity]
  implicit val entityWrites = Json.writes[MyDbEntity]
}

上面的代码无法编译“读取”和“写入”，并显示以下消息：

No unapply function found

将“读取”和“写入”更新为：

  implicit val entityReads: Reads[MyDbEntity] = (
    (__ \ "id").read[Long] and
    (__ \ "field_1").read[String]
    ........
  )(MyDbEntity.apply _)  

  implicit val postWrites: Writes[MyDbEntity] = (
    (__ \ "id").write[Long] and
    (__ \ "user").write[String]
    ........
  )(unlift(MyDbEntity.unapply))

也不起作用：

  implementation restricts functions to 22 parameters

  value unapply is not a member of object models.MyDbEntity

我的理解是 Scala 2.11 在功能上仍然有一些限制，并且像我上面描述的那样的东西尚不可行。 这对我来说似乎很奇怪，因为如果仍然不支持主要用户案例，我看不到取消对案例类的限制的好处，所以我想知道我是否遗漏了什么。

非常欢迎指向问题或实现细节的指针！ 谢谢！

Answer 1

这是不可能的，开箱即用，原因如下：

• 首先，正如gourlaysama指出的那样，play-json 库使用scala 宏来避免样板代码， 当前代码依赖于unapply和apply方法来检索字段。 这解释了您问题中的第一条错误消息。

• 其次，play-json 库依赖于一个函数库，该库当前仅使用与先前案例类字段数量限制相对应的固定数量的参数。 这解释了您问题中的第二条错误消息。

但是，可以通过以下任一方式绕过第二点：

• 使用无形的自动类型类派生功能。 Naveen Gattu写了一个很好的要旨，这样做确实如此。

• 覆盖默认功能构建器

首先，创建缺少的FunctionalBuilder ：

class CustomFunctionalBuilder[M[_]](canBuild: FunctionalCanBuild[M]) extends FunctionalBuilder {

    class CustomCanBuild22[A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22](m1: M[A1 ~ A2 ~ A3 ~ A4 ~ A5 ~ A6 ~ A7 ~ A8 ~ A9 ~ A10 ~ A11 ~ A12 ~ A13 ~ A14 ~ A15 ~ A16 ~ A17 ~ A18 ~ A19 ~ A20 ~ A21], m2: M[A22]) {
def ~[A23](m3: M[A23]) = new CustomCanBuild23[A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22, A23](canBuild(m1, m2), m3)

  def and[A23](m3: M[A23]) = this.~(m3)

  def apply[B](f: (A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22) => B)(implicit fu: Functor[M]): M[B] =
  fu.fmap[A1 ~ A2 ~ A3 ~ A4 ~ A5 ~ A6 ~ A7 ~ A8 ~ A9 ~ A10 ~ A11 ~ A12 ~ A13 ~ A14 ~ A15 ~ A16 ~ A17 ~ A18 ~ A19 ~ A20 ~ A21 ~ A22, B](canBuild(m1, m2), { case a1 ~ a2 ~ a3 ~ a4 ~ a5 ~ a6 ~ a7 ~ a8 ~ a9 ~ a10 ~ a11 ~ a12 ~ a13 ~ a14 ~ a15 ~ a16 ~ a17 ~ a18 ~ a19 ~ a20 ~ a21 ~ a22 => f(a1, a2, a3, a4, a5, a6, a7, a8, a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20, a21, a22) })

  def apply[B](f: B => (A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22))(implicit fu: ContravariantFunctor[M]): M[B] =
  fu.contramap(canBuild(m1, m2), (b: B) => { val (a1, a2, a3, a4, a5, a6, a7, a8, a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20, a21, a22) = f(b); new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(a1, a2), a3), a4), a5), a6), a7), a8), a9), a10), a11), a12), a13), a14), a15), a16), a17), a18), a19), a20), a21), a22) })

  def apply[B](f1: (A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22) => B, f2: B => (A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22))(implicit fu: InvariantFunctor[M]): M[B] =
  fu.inmap[A1 ~ A2 ~ A3 ~ A4 ~ A5 ~ A6 ~ A7 ~ A8 ~ A9 ~ A10 ~ A11 ~ A12 ~ A13 ~ A14 ~ A15 ~ A16 ~ A17 ~ A18 ~ A19 ~ A20 ~ A21 ~ A22, B](
    canBuild(m1, m2), { case a1 ~ a2 ~ a3 ~ a4 ~ a5 ~ a6 ~ a7 ~ a8 ~ a9 ~ a10 ~ a11 ~ a12 ~ a13 ~ a14 ~ a15 ~ a16 ~ a17 ~ a18 ~ a19 ~ a20 ~ a21 ~ a22 => f1(a1, a2, a3, a4, a5, a6, a7, a8, a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20, a21, a22) },
    (b: B) => { val (a1, a2, a3, a4, a5, a6, a7, a8, a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20, a21, a22) = f2(b); new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(new ~(a1, a2), a3), a4), a5), a6), a7), a8), a9), a10), a11), a12), a13), a14), a15), a16), a17), a18), a19), a20), a21), a22) }
  )

  def join[A >: A1](implicit witness1: <:<[A, A1], witness2: <:<[A, A2], witness3: <:<[A, A3], witness4: <:<[A, A4], witness5: <:<[A, A5], witness6: <:<[A, A6], witness7: <:<[A, A7], witness8: <:<[A, A8], witness9: <:<[A, A9], witness10: <:<[A, A10], witness11: <:<[A, A11], witness12: <:<[A, A12], witness13: <:<[A, A13], witness14: <:<[A, A14], witness15: <:<[A, A15], witness16: <:<[A, A16], witness17: <:<[A, A17], witness18: <:<[A, A18], witness19: <:<[A, A19], witness20: <:<[A, A20], witness21: <:<[A, A21], witness22: <:<[A, A22], fu: ContravariantFunctor[M]): M[A] =
  apply[A]((a: A) => (a: A1, a: A2, a: A3, a: A4, a: A5, a: A6, a: A7, a: A8, a: A9, a: A10, a: A11, a: A12, a: A13, a: A14, a: A15, a: A16, a: A17, a: A18, a: A19, a: A20, a: A21, a: A22))(fu)

  def reduce[A >: A1, B](implicit witness1: <:<[A1, A], witness2: <:<[A2, A], witness3: <:<[A3, A], witness4: <:<[A4, A], witness5: <:<[A5, A], witness6: <:<[A6, A], witness7: <:<[A7, A], witness8: <:<[A8, A], witness9: <:<[A9, A], witness10: <:<[A10, A], witness11: <:<[A11, A], witness12: <:<[A12, A], witness13: <:<[A13, A], witness14: <:<[A14, A], witness15: <:<[A15, A], witness16: <:<[A16, A], witness17: <:<[A17, A], witness18: <:<[A18, A], witness19: <:<[A19, A], witness20: <:<[A20, A], witness21: <:<[A21, A], witness22: <:<[A22, A], fu: Functor[M], reducer: Reducer[A, B]): M[B] =
  apply[B]((a1: A1, a2: A2, a3: A3, a4: A4, a5: A5, a6: A6, a7: A7, a8: A8, a9: A9, a10: A10, a11: A11, a12: A12, a13: A13, a14: A14, a15: A15, a16: A16, a17: A17, a18: A18, a19: A19, a20: A20, a21: A21, a22: A22) =>  reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.append(reducer.unit(a1: A), a2: A), a3: A), a4: A), a5: A), a6: A), a7: A), a8: A), a9: A), a10: A), a11: A), a12: A), a13: A), a14: A), a15: A), a16: A), a17: A), a18: A), a19: A), a20: A), a21: A), a22: A))(fu)

  def tupled(implicit v: VariantExtractor[M]): M[(A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22)] =
  v match {
    case FunctorExtractor(fu) => apply { (a1: A1, a2: A2, a3: A3, a4: A4, a5: A5, a6: A6, a7: A7, a8: A8, a9: A9, a10: A10, a11: A11, a12: A12, a13: A13, a14: A14, a15: A15, a16: A16, a17: A17, a18: A18, a19: A19, a20: A20, a21: A21, a22: A22) => (a1, a2, a3, a4, a5, a6, a7, a8, a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20, a21, a22) }(fu)
    case ContravariantFunctorExtractor(fu) => apply[(A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22)] { (a: (A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22)) => (a._1, a._2, a._3, a._4, a._5, a._6, a._7, a._8, a._9, a._10, a._11, a._12, a._13, a._14, a._15, a._16, a._17, a._18, a._19, a._20, a._21, a._22) }(fu)
    case InvariantFunctorExtractor(fu) => apply[(A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22)]({ (a1: A1, a2: A2, a3: A3, a4: A4, a5: A5, a6: A6, a7: A7, a8: A8, a9: A9, a10: A10, a11: A11, a12: A12, a13: A13, a14: A14, a15: A15, a16: A16, a17: A17, a18: A18, a19: A19, a20: A20, a21: A21, a22: A22) => (a1, a2, a3, a4, a5, a6, a7, a8, a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20, a21, a22) }, { (a: (A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22)) => (a._1, a._2, a._3, a._4, a._5, a._6, a._7, a._8, a._9, a._10, a._11, a._12, a._13, a._14, a._15, a._16, a._17, a._18, a._19, a._20, a._21, a._22) })(fu)
    }

  }

  class CustomCanBuild23[A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22, A23](m1: M[A1 ~ A2 ~ A3 ~ A4 ~ A5 ~ A6 ~ A7 ~ A8 ~ A9 ~ A10 ~ A11 ~ A12 ~ A13 ~ A14 ~ A15 ~ A16 ~ A17 ~ A18 ~ A19 ~ A20 ~ A21 ~ A22], m2: M[A23]) {
  }

}

然后通过提供您自己的FunctionalBuilderOps实例：

implicit def customToFunctionalBuilderOps[M[_], A](a: M[A])(implicit fcb: FunctionalCanBuild[M]) = new CustomFunctionalBuilderOps[M, A](a)(fcb)

最后，关于第一点，我已经发送了一个pull request来尝试简化当前的实现。

Answer 2

我们还将模型分解为多个案例类，但这很快变得难以管理。 我们使用Slick作为我们的对象关系映射器，并且 Slick 2.0 带有一个代码生成器，我们用来生成类（带有应用方法和复制构造函数来模拟案例类）以及从 Json 实例化模型的方法（我们不会自动生成将模型转换为 Json 的方法，因为我们有太多特殊情况需要处理）。 使用 Slick 代码生成器不需要您使用 Slick 作为对象关系映射器。

这是代码生成器输入的一部分 - 此方法采用 JsObject 并使用它来实例化新模型或更新现有模型。

private def getItem(original: Option[${name}], json: JsObject, trackingData: TrackingData)(implicit session: scala.slick.session.Session): Try[${name}] = {
  preProcess("$name", columnSet, json, trackingData).flatMap(updatedJson => {
    ${indent(indent(indent(entityColumnsSansId.map(c => s"""val ${c.name}_Parsed = parseJsonField[${c.exposedType}](original.map(_.${c.name}), "${c.name}", updatedJson, "${c.exposedType}")""").mkString("\n"))))}
    val errs = Seq(${indent(indent(indent(indent(entityColumnsSansId.map(c => s"${c.name}_Parsed.map(_ => ())").mkString(", ")))))}).condenseUnit
    for {
      _ <- errs
      ${indent(indent(indent(indent(entityColumnsSansId.map(c => s"${c.name}_Val <- ${c.name}_Parsed").mkString("\n")))))}
    } yield {
      original.map(_.copy(${entityColumnsSansId.map(c => s"${c.name} = ${c.name}_Val").mkString(", ")}))
        .getOrElse(${name}.apply(id = None, ${entityColumnsSansId.map(c => s"${c.name} = ${c.name}_Val").mkString(", ")}))
    }
  })
}

例如，使用我们的 ActivityLog 模型，这会生成以下代码。 如果“original”是None，那么它是从“createFromJson”方法调用的，我们实例化一个新模型； 如果“原始”是 Some(activityLog) 则从“updateFromJson”方法调用它，我们更新现有模型。 在“val errs = ...”行上调用的“condenseUnit”方法接受一个 Seq[Try[Unit]] 并产生一个 Try[Unit]； 如果 Seq 有任何错误，则 Try[Unit] 连接异常消息。 parseJsonField 和 parseField 方法没有生成——它们只是从生成的代码中引用。

private def parseField[T](name: String, json: JsObject, tpe: String)(implicit r: Reads[T]): Try[T] = {
  Try((json \ name).as[T]).recoverWith {
    case e: Exception => Failure(new IllegalArgumentException("Failed to parse " + Json.stringify(json \ name) + " as " + name + " : " + tpe))
  }
}

def parseJsonField[T](default: Option[T], name: String, json: JsObject, tpe: String)(implicit r: Reads[T]): Try[T] = {
  default match {
    case Some(t) => if(json.keys.contains(name)) parseField(name, json, tpe)(r) else Try(t)
    case _ => parseField(name, json, tpe)(r)
  }
}

private def getItem(original: Option[ActivityLog], json: JsObject, trackingData: TrackingData)(implicit session: scala.slick.session.Session): Try[ActivityLog] = {
  preProcess("ActivityLog", columnSet, json, trackingData).flatMap(updatedJson => {
    val user_id_Parsed = parseJsonField[Option[Int]](original.map(_.user_id), "user_id", updatedJson, "Option[Int]")
    val user_name_Parsed = parseJsonField[Option[String]](original.map(_.user_name), "user_name", updatedJson, "Option[String]")
    val item_id_Parsed = parseJsonField[Option[String]](original.map(_.item_id), "item_id", updatedJson, "Option[String]")
    val item_item_type_Parsed = parseJsonField[Option[String]](original.map(_.item_item_type), "item_item_type", updatedJson, "Option[String]")
    val item_name_Parsed = parseJsonField[Option[String]](original.map(_.item_name), "item_name", updatedJson, "Option[String]")
    val modified_Parsed = parseJsonField[Option[String]](original.map(_.modified), "modified", updatedJson, "Option[String]")
    val action_name_Parsed = parseJsonField[Option[String]](original.map(_.action_name), "action_name", updatedJson, "Option[String]")
    val remote_ip_Parsed = parseJsonField[Option[String]](original.map(_.remote_ip), "remote_ip", updatedJson, "Option[String]")
    val item_key_Parsed = parseJsonField[Option[String]](original.map(_.item_key), "item_key", updatedJson, "Option[String]")
    val created_at_Parsed = parseJsonField[Option[java.sql.Timestamp]](original.map(_.created_at), "created_at", updatedJson, "Option[java.sql.Timestamp]")
    val as_of_date_Parsed = parseJsonField[Option[java.sql.Timestamp]](original.map(_.as_of_date), "as_of_date", updatedJson, "Option[java.sql.Timestamp]")
    val errs = Seq(user_id_Parsed.map(_ => ()), user_name_Parsed.map(_ => ()), item_id_Parsed.map(_ => ()), item_item_type_Parsed.map(_ => ()), item_name_Parsed.map(_ => ()), modified_Parsed.map(_ => ()), action_name_Parsed.map(_ => ()), remote_ip_Parsed.map(_ => ()), item_key_Parsed.map(_ => ()), created_at_Parsed.map(_ => ()), as_of_date_Parsed.map(_ => ())).condenseUnit
    for {
      _ <- errs
      user_id_Val <- user_id_Parsed
      user_name_Val <- user_name_Parsed
      item_id_Val <- item_id_Parsed
      item_item_type_Val <- item_item_type_Parsed
      item_name_Val <- item_name_Parsed
      modified_Val <- modified_Parsed
      action_name_Val <- action_name_Parsed
      remote_ip_Val <- remote_ip_Parsed
      item_key_Val <- item_key_Parsed
      created_at_Val <- created_at_Parsed
      as_of_date_Val <- as_of_date_Parsed
    } yield {
      original.map(_.copy(user_id = user_id_Val, user_name = user_name_Val, item_id = item_id_Val, item_item_type = item_item_type_Val, item_name = item_name_Val, modified = modified_Val, action_name = action_name_Val, remote_ip = remote_ip_Val, item_key = item_key_Val, created_at = created_at_Val, as_of_date = as_of_date_Val))
        .getOrElse(ActivityLog.apply(id = None, user_id = user_id_Val, user_name = user_name_Val, item_id = item_id_Val, item_item_type = item_item_type_Val, item_name = item_name_Val, modified = modified_Val, action_name = action_name_Val, remote_ip = remote_ip_Val, item_key = item_key_Val, created_at = created_at_Val, as_of_date = as_of_date_Val))
    }
  })
}

Answer 3

您可以使用 Jackson 的 Scala 模块。 Play 的 json 功能建立在 Jackson scala 之上。 我不知道为什么他们在这里设置了 22 个字段的限制，而 jackson 支持超过 22 个字段。 一个函数调用永远不能使用超过 22 个参数可能是有道理的，但是我们可以在一个 DB 实体中有数百个列，所以这里的这个限制是荒谬的，并且使 Play 成为一个效率较低的玩具。 看一下这个：

import com.fasterxml.jackson.databind.ObjectMapper
import com.fasterxml.jackson.module.scala.experimental.ScalaObjectMapper
import com.fasterxml.jackson.module.scala.DefaultScalaModule

object JacksonUtil extends App {
  val mapper = new ObjectMapper with ScalaObjectMapper
  mapper.registerModule(DefaultScalaModule)


  val t23 = T23("a","b","c","d","e","f","g","h","i","j","k","l","m","n","o","p","q","r","s","t","u","v","w")

  println(mapper.writeValueAsString(t23))
 }
case class T23(f1:String,f2:String,f3:String,f4:String,f5:String,f6:String,f7:String,
    f8:String,f9:String,f10:String,f11:String,f12:String,f13:String,f14:String,f15:String,
    f16:String,f17:String,f18:String,f19:String,f20:String,f21:String,f22:String,f23:String)

Answer 4

这似乎很好地处理了这一切。

+22 字段案例类格式化程序以及更多用于 play-json https://github.com/xdotai/play-json-extensions

支持 Scala 2.11.x、2.12.x 和 2.13.x 并播放 2.3、2.4、2.5 和 2.7

并且在play-json 问题中被引用为首选解决方案（但尚未合并）

Answer 5

我正在制作一个图书馆。 请试试这个https://github.com/xuwei-k/play-twenty-three

Answer 6

案例类可能不起作用的情况； 其中一种情况是案例类不能超过 22 个字段。 另一种情况可能是您事先不了解架构。 在这种方法中，数据作为行对象的 RDD 加载。 Schema 是使用 StructType 和 StructField 对象分别创建的，它们分别代表一个表和一个字段。 Schema 应用于行 RDD 以在 Spark 中创建DataFrame 。

Answer 7

我尝试了另一个答案中提出的基于 Shapeless“Automatic Typeclass Derivation”的解决方案，但它对我们的模型不起作用 - 抛出 StackOverflow 异常（具有约 30 个字段的案例类和具有 4-10 个字段的案例类的 4 个嵌套集合）。

所以，我们采用了这个解决方案，它完美地工作。 通过编写 ScalaCheck 测试确认了这一点。 请注意，它需要 Play Json 2.4。

Answer 8

在 dotty (Scala 3) 中，您现在可以在 Case 类中使用超过 22 个字段。