如何将动态类型传递给 scala 中已经具有泛型类型的方法

Question

I have a method:

object Object1{
def method1[T >: Null: ClassTag: TypeTag](input: String): T = {
    val m = new ObjectMapper() with ScalaObjectMapper
    m.registerModule(DefaultScalaModule)
    Option(input)
      .map(m.readValue[T](_, classTag[T].runtimeClass.asInstanceOf[Class[T]]))
      .orNull
  }
}

I want to dynamically pass type to this method, I have tried it like:

import spark.implicits._

def getHome[T](s: Dataset[Sig], someString: String): Dataset[T] = {
    s.filter(s => (s.sType == someString))
      .map(s => Object1.method1[T](s.sDetails))
  }

But I am getting error: No implicits found for parameter evidence$6 . I have also imported implicits. What is the best way to implement above logic in Scala?

Answer 1

You are trying to pass a value of a wider type T to a narrower type T >: Null: ClassTag: TypeTag and this is never OK. You have to add the same (or better) type constraints to the type parameter of getHome :

def getHome[T >: Null: ClassTag: TypeTag](s: Dataset[Sig], someString: String): Dataset[T] = {

It also feels like method1 should be using the Manifest typeclass rather than ClassTag and TypeTag . This might simplify things, so check the docs for ScalaObjectMapper .