Scala 2.11 + Play Framework 2.3 案例类和函数中的 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 Issue、Release Notes）。

这对我来说一直是个问题，因为我使用案例类对 Play + Postgres Async 中具有超过 22 个字段的数据库实体进行建模。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很难维护和扩展，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]
}

上面的代码编译失败，“读取”和“写入”都显示以下消息：

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))

也不行：

  implementation restricts functions to 22 parameters

  value unapply is not a member of object models.MyDbEntity

我的理解是 Scala 2.11 在功能上仍然存在一些限制，并且像我上面描述的那样还不可能。这对我来说似乎很奇怪，因为如果一个主要用户案例仍然不受支持，我看不到解除案例类限制的好处，所以我想知道我是否遗漏了什么。

非常欢迎指出问题或实施细节！谢谢！

Answer 1

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

Answer 2

看起来这一切都很好。

+22 字段案例类格式化程序和更多用于 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 issue 中被引用为首选解决方案（但尚未合并）

Answer 3

我尝试了另一个答案中提出的基于 Shapeless “自动类型类派生”的解决方案，但它不适用于我们的模型 - 引发 StackOverflow 异常（具有约 30 个字段的案例类和 4 个具有 4-10 的案例类的嵌套集合字段）。

所以，我们采用了this 解决方案，它完美无缺。通过编写 ScalaCheck 测试确认了这一点。请注意，它需要 Play Json 2.4。

Answer 4

案例类可能不起作用的案例；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模式。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以创建DataFrame in Spark。

Answer 5

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 个字段限制，而杰克逊支持超过 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 实体中拥有数百个列，所以这里的这个限制是荒谬的，并且使 Play 成为一个生产力较低的玩具。 看看这个：

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 6

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

Answer 7

这是不可能的，开箱即用，有几个原因：

• 首先，正如gourlaysama 指出的那样，play-json 库使用了scala macro to avoid bolierplate code，而current code 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 functional library，它目前仅适用于 with a fixed number of parameters，对应于之前的案例类字段数量限制。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shapeless 自动类型类派生功能。 Naveen Gattu 写了 an excellent gist 就是这样做的。

• 覆盖默认功能构建器

首先，创建缺失的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，我发了pull request 来尝试简化当前的实现。

Answer 8

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，Slick 2.0 带有一个code generator，我们用它来生成类（它带有应用方法和复制构造函数来模仿案例类）以及从 Json 实例化模型的方法（我们不要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 模型会生成以下代码。如果“original”是None，那么这是从“createFromJson”方法调用的，我们实例化一个新模型；如果“original”是 Some(activityLog)，那么这是从“updateFromJson”方法调用的，我们更新现有模型。在“val errs = ...”行上调用的“condenseUnit”方法需要一个 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。不会生成 parseJsonField 和 parseField 方法 - 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))
    }
  })
}