自定义映射到 Slick 中的嵌套案例 class 结构（超过 22 列）

Question

我正在尝试将超过 22 列的数据库行映射到案例 class 树。我宁愿不使用 HList，因为我不想使用它 API，还有我在某处读到的一些指数编译时间反馈...

我已经阅读了 Stefan Zeiger 回答的这个帖子：How can I handle a > 22 column table with Slick using nested tuples or HLists?

我看过这个演示如何定义自定义映射函数的测试，我想这样做：

https://github.com/slick/slick/blob/2.1/slick-testkit/src/main/scala/com/typesafe/slick/testkit/tests/JdbcMapperTest.scala#L129-141

def * = (
        id,
        (p1i1, p1i2, p1i3, p1i4, p1i5, p1i6),
        (p2i1, p2i2, p2i3, p2i4, p2i5, p2i6),
        (p3i1, p3i2, p3i3, p3i4, p3i5, p3i6),
        (p4i1, p4i2, p4i3, p4i4, p4i5, p4i6)
      ).shaped <> ({ case (id, p1, p2, p3, p4) =>
        // We could do this without .shaped but then we'd have to write a type annotation for the parameters
        Whole(id, Part.tupled.apply(p1), Part.tupled.apply(p2), Part.tupled.apply(p3), Part.tupled.apply(p4))
      }, { w: Whole =>
        def f(p: Part) = Part.unapply(p).get
        Some((w.id, f(w.p1), f(w.p2), f(w.p3), f(w.p4)))
      })

问题是我做不到！

我试过更小的步骤。

class UserTable(tag: Tag) extends TableWithId[User](tag,"USER") {
  override def id = column[String]("id", O.PrimaryKey)
  def role = column[UserRole.Value]("role", O.NotNull)
  def login = column[String]("login", O.NotNull)
  def password = column[String]("password", O.NotNull)
  def firstName = column[String]("first_name", O.NotNull)
  def lastName = column[String]("last_name", O.NotNull)
  //
  def * = (id, role, login, password, firstName, lastName) <> (User.tupled,User.unapply)
  //
  def login_index = index("idx_user_login", login, unique = true)
}

好像我打电话的时候

(id, (firstName, lastName)).shaped

类型是ShapedValue[(Column[String], (Column[String], Column[String])), Nothing]

虽然这个似乎工作正常

(id, firstName, lastName).shaped

U型参数不是Nothing而是符合预期(String, String, String)

我不太了解 Slick 的所有内部结构是如何工作的。有人可以解释为什么我不能让我的代码工作吗？是否缺少导入或其他内容？

我想我需要得到一个

类型的值

ShapedValue[(Column[String], (Column[String], Column[String])), (String, (String, String))]

但我不知道为什么 returns 我 Nothing 也不太明白这些隐式 Shape 参数来自哪里...

我想要的只是能够轻松地将我的专栏分成 2 个案例 classes

谢谢

Answer 1

此外，22 列限制也有同样的问题，test case 很有帮助。不确定为什么示例代码不适合你，下面的代码对我来说很好，

case class UserRole(var role: String, var extra: String)
case class UserInfo(var login: String, var password: String, var firstName: String, var lastName: String)

case class User(id: Option[String], var info: UserInfo, var role: UserRole)

class UserTable(tag: Tag) extends Table[User](tag, "USER") {

  def id = column[String]("id", O.PrimaryKey)
  def role = column[String]("role", O.NotNull)
  def extra = column[String]("extra", O.NotNull)
  def login = column[String]("login", O.NotNull)
  def password = column[String]("password", O.NotNull)
  def firstName = column[String]("first_name", O.NotNull)
  def lastName = column[String]("last_name", O.NotNull)

  /** Projection */
  def * = (
    id,
    (login, password, firstName, lastName),
    (role, extra)
  ).shaped <> (

  { case (id, userInfo, userRole) =>
    User(Option[id], UserInfo.tupled.apply(userInfo), UserRole.tupled.apply(userRole))
  },
  { u: User =>
      def f1(p: UserInfo) = UserInfo.unapply(p).get
      def f2(p: UserRole) = UserRole.unapply(p).get
      Some((u.id.get, f1(u.info), f2(u.role)))
  })

  def login_index = index("id_user_login", login, unique = true)
}

Answer 2

正如 Izongren 所回答的那样，它工作正常，但是很难编写这样的代码，因为使用很长的元组很烦人...我决定这样做 "step by step" 并始终明确提供类型以避免类型推断问题，现在它工作正常。

case class Patient(
                    id: String = java.util.UUID.randomUUID().toString,
                    companyScopeId: String,
                    assignedToUserId: Option[String] = None,
                    info: PatientInfo
                    ) extends ModelWithId



case class PatientInfo(
                        firstName: Option[String] = None,
                        lastName: Option[String] = None,
                        gender: Option[Gender.Value] = None,
                        alias: Option[String] = None,
                        street: Option[String] = None,
                        city: Option[String] = None,
                        postalCode: Option[String] = None,
                        phone: Option[String] = None,
                        mobilePhone: Option[String] = None,
                        email: Option[String] = None,
                        age: Option[AgeRange.Value] = None,
                        companySeniority: Option[CompanySeniorityRange.Value] = None,
                        employmentContract: Option[EmploymentContract.Value] = None,
                        socialStatus: Option[SocialStatus.Value] = None,
                        jobTitle: Option[String] = None
                        )

class PatientTable(tag: Tag) extends TableWithId[Patient](tag,"PATIENT") {
  override def id = column[String]("id", O.PrimaryKey)
  def companyScopeId = column[String]("company_scope_id", O.NotNull)
  def assignedToUserId = column[Option[String]]("assigned_to_user_id", O.Nullable)

  def firstName = column[Option[String]]("first_name", O.Nullable)
  def lastName = column[Option[String]]("last_name", O.Nullable)
  def gender = column[Option[Gender.Value]]("gender", O.Nullable)
  def alias = column[Option[String]]("alias", O.Nullable)
  def street = column[Option[String]]("street", O.Nullable)
  def city = column[Option[String]]("city", O.Nullable)
  def postalCode = column[Option[String]]("postal_code", O.Nullable)
  def phone = column[Option[String]]("phone", O.Nullable)
  def mobilePhone = column[Option[String]]("mobile_phone", O.Nullable)
  def email = column[Option[String]]("email", O.Nullable)
  def age = column[Option[AgeRange.Value]]("age", O.Nullable)
  def companySeniority = column[Option[CompanySeniorityRange.Value]]("company_seniority", O.Nullable)
  def employmentContract = column[Option[EmploymentContract.Value]]("employment_contract", O.Nullable)
  def socialStatus = column[Option[SocialStatus.Value]]("social_status", O.Nullable)
  def jobTitle = column[Option[String]]("job_title", O.Nullable)
  def role = column[Option[String]]("role", O.Nullable)



  private type PatientInfoTupleType = (Option[String], Option[String], Option[Gender.Value], Option[String], Option[String], Option[String], Option[String], Option[String], Option[String], Option[String], Option[AgeRange.Value], Option[CompanySeniorityRange.Value], Option[EmploymentContract.Value], Option[SocialStatus.Value], Option[String])
  private type PatientTupleType = (String, String, Option[String], PatientInfoTupleType)
  //
  private val patientShapedValue = (id, companyScopeId, assignedToUserId,
    (
      firstName, lastName, gender, alias, street, city, postalCode,
      phone, mobilePhone,email, age, companySeniority, employmentContract, socialStatus, jobTitle
      )
    ).shaped[PatientTupleType]
  //
  private val toModel: PatientTupleType => Patient = { patientTuple =>
    Patient(
      id = patientTuple._1,
      companyScopeId = patientTuple._2,
      assignedToUserId = patientTuple._3,
      info = PatientInfo.tupled.apply(patientTuple._4)
    )
  }
  private val toTuple: Patient => Option[PatientTupleType] = { patient =>
    Some {
      (
        patient.id,
        patient.companyScopeId,
        patient.assignedToUserId,
        (PatientInfo.unapply(patient.info).get)
        )
    }
  }

  def * = patientShapedValue <> (toModel,toTuple)
}

Answer 3

另外，你也可以这样使用

class UserTable(tag: Tag) extends TableWithId[User](tag,"USER") {
  def id = column[String]("id", O.PrimaryKey)
  def role = column[String]("role", O.NotNull)
  def extra = column[String]("extra", O.NotNull)
  def login = column[String]("login", O.NotNull)
  def password = column[String]("password", O.NotNull)
  def firstName = column[String]("first_name", O.NotNull)
  def lastName = column[String]("last_name", O.NotNull)

  def loginMap = (login, password, firstName, lastName) <> (UserInfo.tupled, UserInfo.unapply)

  def roleMap = (role, extra) <> (Role.tupled, Role.unapply)

  override def * = (id, roleMap, loginMap) <> (User.tupled,User.unapply)

  def login_index = index("idx_user_login", login, unique = true)
}

自定义映射到 Slick 中的嵌套案例 class 结构（超过 22 列）

Custom mapping to nested case class structure in Slick (more than 22 columns)

scala

slick