无法使用 SkinnyORM 在模型的嵌套对象中检索嵌套对象列表

Cannot retrieve nested object list within nested object in model with SkinnyORM

我有一个模型 Contact 与模型 Role 有 hasMany 关系。 Model Role 与 Phone 具有 hasManyThrough 关系,中间 table RolePhone。 所有检索操作都设置为急切。

当我检索联系人时,我看到了它的角色列表,但每个角色都有一个空的 Phone 列表。但是当我检索所有角色时,他们的 Phone 列表包含正确的 Phones.

联系人和角色中的关系都被标记为默认值,我尝试手动设置角色的 joinDefinitions enabledEvenIfAssociated = true 但这些都不起作用

这些是我的模型,仅显示结构和关系

case class Contact(id: Long, name: String, roles: Seq[Role] = Nil)

object Contact extends SkinnyCRUDMapper[Contact] {
  hasMany[Role](
    Role -> Role.defaultAlias,
    (contact, role) => sqls.eq(contact.id, role.contactId),
    (contact, roles) => contact.copy(roles = roles),
  ).byDefault
}

case class RoleCommand(name: String, phoneNumbers: Seq[Int])

case class Role(id: Long, name: String, contactId: Long, phones: Seq[Phone] = Nil)

object Role extends SkinnyCRUDMapper[Role] {
  hasManyThrough[RolePhone, Phone](
    through = RolePhone -> RolePhone.defaultAlias,
    throughOn = (r, rp) => sqls.eq(r.id, rp.roleId),
    many = Phone -> Phone.defaultAlias,
    on = (rp, p) => sqls.eq(rp.phoneId, p.id),
    merge = (role, phones) => role.copy(phones = phones)
  ).byDefault
}

case class RolePhone(roleId: Long, phoneId: Long)
case class Phone(id: Long, number: Int)

这是我的数据库结构

create table contact (
  id bigint(20) not null auto_increment primary key,
  name varchar(50) not null unique
) engine=innoDB;

create table role (
  id bigint(20) not null auto_increment primary key,
  name varchar(50) not null unique,
  contact_id bigint(20) not null
) engine=innoDB;

create table phone (
  id bigint(20) not null auto_increment primary key,
  number int(20) not null
) engine=innoDB;

create table role_phone (
  role_id bigint(20) not null,
  phone_id bigint(20) not null
) engine=innoDB;

这是一个 Main class 添加以暴露问题

object Main extends App with Connection with DatasourceConfig  {

  val contact1 = Contact.save("contact1",
    Seq(
      RoleCommand("role1", Seq(12345, 12321)),
      RoleCommand("role2", Seq(54321, 12321)),
      RoleCommand("role3", Seq(12345, 54321)),
    )
  )

  // retrieving a contact won't fill the phones number inside each role
  println(s"My Contact: ${Contact.findById(contact1)}")

  // retrieving a role will fill the phones list
  println(s"All Roles: ${Role.findAll()}")

}

完整代码位于 https://github.com/airabinovich/testing_skinny

我预计检索联系人会得到每个角色中的电话列表 按照上面的例子,应该是这样的

Contact(
  1,
  contact1,
  Vector(
    Role(1,role1,1,Vector(Phone(1,12345), Phone(2,12321))),
    Role(2,role2,1,Vector(Phone(3,54321), Phone(2,12321))),
    Role(3,role3,1,Vector(Phone(1,12345), Phone(3,54321)))
  )
)

但我明白了

Contact(
  1,
  contact1,
  Vector(
    Role(1,role1,1,List()),
    Role(1,role1,1,List()),
    Role(2,role2,1,List()),
    Role(2,role2,1,List()),
    Role(3,role3,1,List()),
    Role(3,role3,1,List())
  )
)

Kazuhiro Sera answered in his pull request一样,可以通过稍微改变一下关系来完成预先加载

object Contact extends SkinnyCRUDMapper[Contact] {
  val rolesRef = hasMany[Role](
    Role -> Role.defaultAlias,
    (contact, role) => sqls.eq(contact.id, role.contactId),
    (contact, roles) => contact.copy(roles = roles),
  ).includes[Role](
    merge = (contacts, roles) => contacts.map(c => c.copy(roles = roles.filter(_.contactId == c.id)))
   ).byDefault
}

并且在从数据库中检索时包括关系

 println(s"My Contact: ${Contact.includes(Contact.rolesRef).findById(contact1)}")