Dynamic member lookup in Scala

Dynamic member lookup in Scala

Dynamic member lookup in Scala

A lot of dynamic typed languages has function that catch all messages the target object doesn’t have. It gives these languages a lot of power to write DSL.

Scala 2.9 introduced an experimental option that enables this sort of dynamic handling of accesses to types in ways that would otherwise fail static type checking. It was refined and made non-experimental in 2.10, though it is still controlled through flag which is disabled by default.

Dynamic trait

Scala’s Dynamic type allows you to access instance members that aren’t statically defined in a class. It’s a marker trait with no members – it provides a hint to the compiler to rewrite method calls and field accesses to missing members as calls to one of the following methods:

  • selectDynamic – method allows to write field accessors
  • updateDynamic – method allows to write field updates
  • applyDynamic – method allows to call methods with arguments
  • applyDynamicNamed – method allows to call methods with named arguments

To shed more light on how these methods might work see examples below together with overview of dynamic translations that are done under the hood:

dyn.foo                     //dyn.selectDynamic("foo")
dyn.foo = "bar"             //dyn.updateDynamic("foo")("bar")  
dyn.foo("bar")              //dyn.applyDynamic("foo")("bar")  
dyn.foo(name = "bar")       //dyn.applyDynamicNamed("foo")(("name" -> "bar"))
dyn.foo(5, name = "bar")    //dyn.applyDynamicNamed("foo")(("", 5),("name", "bar"))  
dyn.foo("bar") = "rab"      //dyn.applyDynamic("foo").update("bar", "rab")

So what you should do if you want to add dynamic member lookup in your code? It’s simple. You just need to write a class that extends Dynamic trait. Inside this class you can implement any of these four methods.

But that is not all. As I mentioned above it’s disabled feature by default so you need to set the compiler option -language:dynamics or add a import scala.language.dynamics to make the feature work.

And that’s all you need to do to make your statically typed code more dynamic.

Sample implementation

As an example, we might want to build our own Map implementation. Our Map needs to do few things: – store data as key-value, where key is String and value is Generic type, – return default(defined) value for keys that doesn’t exists.

We can do it like this :

class MyMap[T](defaultValue: T) extends Dynamic {

  val values: mutable.Map[String, T] = mutable.Map.empty[String, T]

  def selectDynamic(name: String): T = {
    values.get(name) match {
      case Some(value) => value
      case None => defaultValue
    }
  }

  def updateDynamic(name: String)(value: T): Unit = {
    values(name) = value
  }

}

So now we can use it in this way:

val myKeyStore = new MyMap(150)

myKeyStore.foo           //Int: 150
myKeyStore.foo = 69
myKeyStore.foo           //Int: 69

How it works? I think, the code above is self-explanatory. Every time a message is sent to this object, it will try to find value with the key of a sent message in the included Map. If it doesn’t find any, it will return the value defined as default.

Going further

With Dynamic you can also create your own custom builders…

case class Ingredient(name: String, amount: Int = 1)

case class Drink(ingredients: List[Ingredient])

case class Shaker(ingredients: List[Ingredient], multiplier: Int = 1) extends Dynamic {

    def selectDynamic(name: String) = {
      name match {
        case c if c.matches("with|and") => this
        case double if double.matches("double|twice|two") => copy(multiplier = 2)
        case triple if triple.matches("triple|three") => copy(multiplier = 3)
        case ingredient => copy(ingredients :+ Ingredient(ingredient, multiplier), multiplier = 1)
      }
    }

    def please = Drink(ingredients)

}

object Shaker {
    def apply(): Shaker = {
      new Shaker(List.empty[Ingredient], 1)

Now you can build you own drinks in easy way:

Shaker().gin.triple.vodka.twice.lemon.and.ice.please
//result: Drink(List(Ingredient(gin,1), Ingredient(vodka,3), Ingredient(lemon,2), Ingredient(ice,1)))

Dynamic and implicit conversions

Implicit conversions are applied before missing methods are passed into updateDynamicselectDynamicapplyDynamic or applyDynamicNamedmethod. To prove that, Lets go back to first example.

When we call key that doesn’t exists, the MyMap return the value 150.

myKeyStore.ten  		//Int: 150

Let’s see what happens with implicit conversion in scope. First we define implicit class:

import scala.language.implicitConversions

object DynamicImplicit {

  implicit class Ten(obj: Any) {
    def ten = 10
  }
}

Import it in the file where we use our map, and try again to call method above.

import DynamicImplicit._
  
myKeyStore.ten				//Int: 10

Dynamic and postfixOps

Sadly, it does not work well together. Compiler doesn’t know which method it should apply if you chain methods as in the example above. So, you have to use dot as methods separator.

More

You can read about Dynamic in the SIP17. If you want to read more about “modularization” of Scala 2.10 and feature flags you should look into SIP18

@marioosh

Do you like this post? Want to stay updated? Follow us on Twitter or subscribe to our Feed.

See also:

Download e-book:

Scalac Case Study Book

Download now

Authors

Mariusz Nosiński

I’m an experienced developer who has acquired a broad knowledge. I’m always ready for new challenges and learning new skills.

Latest Blogposts

29.04.2024 / By  Matylda Kamińska

Scalendar May 2024

scalendar may 2024

Event-driven Newsletter Welcome to our May 2024 edition of Scalendar! As we move into a bustling spring, this issue brings you a compilation of the most anticipated frontend and software architecture events around the globe. With a particular focus on Scala conferences in May 2024, our newsletter serves as your guide to keep you updated […]

23.04.2024 / By  Bartosz Budnik

Kalix tutorial: Building invoice application

Kalix app building.

Scala is well-known for its great functional scala libraries which enable the building of complex applications designed for streaming data or providing reliable solutions with effect systems. However, there are not that many solutions which we could call frameworks to provide every necessary tool and out-of-the box integrations with databases, message brokers, etc. In 2022, Kalix was […]

17.04.2024 / By  Michał Szajkowski

Mocking Libraries can be your doom

Test Automations

Test automation is great. Nowadays, it’s become a crucial part of basically any software development process. And at the unit test level it is often a necessity to mimic a foreign service or other dependencies you want to isolate from. So in such a case, using a mock library should be an obvious choice that […]

software product development

Need a successful project?

Estimate project