这是我的示例scala代码:
object App {
abstract class BaseAction
type ApiAction[T <: BaseAction] = (T) => Unit
case class FirstAction(name: String) extends BaseAction
case class SecondAction(surname: String) extends BaseAction
def action1[Z <: BaseAction] = {
(a: Z) => { // Here i'would like to have a: FirstAction
val z = a.asInstanceOf[FirstAction]
println("Running action: " + z.name )
}
}
def action2[Z <: BaseAction] = {
(a: Z) => { // Here i'would like to have a: SecondAction
val z = a.asInstanceOf[SecondAction]
println("Running action " + z.surname )
}
}
def myActions[T <: BaseAction] = Map[String, ApiAction[T]]("a1" -> action1[T], "a2" -> action2[T])
myActions("a1")(FirstAction("Action 1"))
myActions("a2")(SecondAction("Action 2"))
}
我有很少的动作功能,它们做了不同的事情。 每个动作函数都接收一个参数:action class,其中所有动作类都继承自BaseAction抽象类。
功能myActions是actionName到action函数的映射。
我的代码正在运行,但我认为使用asInstanceOf不是一个好习惯,我想知道如何只使用泛型类型编写此代码,而不使用asInstanceOf。
答案 0 :(得分:0)
问题是你做了很多事情&#34;你不应该做这些事情&#34;。
我会尽力给你更好的&#34; (根据大多数斯卡拉人)写同样的事情的方式。这些更改包括使用来自类型边界的信息以及自定义类型ApiAction
来编写更可预测和有组织的代码。
首先,你有以下抽象,
abstract class BaseAction
type ApiAction[T <: BaseAction] = (T) => Unit
case class FirstAction(name: String) extends BaseAction
case class SecondAction(surname: String) extends BaseAction
现在您可以使用这些抽象来编写Actions
对象,
object MyActions {
val action1: ApiAction[FirstAction] = {
case FirstAction(name) => println("Running action :: " + name)
}
val action1Other: ApiAction[FirstAction] = (fa: FirstAction) => {
println("Running action :: " + fa.name)
}
val action2: ApiAction[SecondAction] = {
case SecondAction(surname) => println("Running action :: " + surname)
}
val action2Other: ApiAction[SecondAction] = (sa: SecondAction) => {
println("Running action :: " + sa.surname)
}
// but lets say you wanted a generic ApiAction
val actionGeneric: ApiAction[BaseAction] = {
case FirstAction(name) => println("Running action :: " + name)
case SecondAction(surname) => println("Running action :: " + surname)
}
}
现在你可以使用这些&#34;动作&#34;如你所愿,在你的应用程序中,
object MyApp extends App {
MyActions.action1(FirstAction("Action 1"))
MyActions.action1Other(FirstAction("Action 1 Other"))
MyActions.actionGeneric(FirstAction("Action 1 Generic"))
MyActions.action2(SecondAction("Action 2"))
MyActions.action2Other(SecondAction("Action 2 Other"))
MyActions.actionGeneric(SecondAction("Action 2 Generic"))
}
答案 1 :(得分:0)
您可以使用惯用的Scala模式匹配而不是asInstanceOf
:
val z = a match {
case FirstAction(name) => println("Running action " + name)
case _ => println("Error")
}
请注意,模式匹配仍然使用isInstanceof
+ asInstanceOf
,但与直接调用asInstanceOf
不同,它被视为良好做法。
顺便说一下,组织你的代码可能是明智的,这样你只需匹配一次,而不是有两个单独的“第一个动作或错误”和“第二个动作或错误”块:
def action[Z <: BaseAction] = {
(a: Z) => a match {
case FirstAction(name) => println("Running action " + name)
case SecondAction(surname) => println("Running action " + surname)
case _ => println("Error")
}
}
def myActions[T <: BaseAction] = Map[String, ApiAction[T]]("a1" -> action[T], "a2" -> action[T])
myActions("a1")(FirstAction("Action 1"))
myActions("a2")(SecondAction("Action 2"))
// output:
// Running action Action 1
// Running action Action 2