Limits of Nat type in Shapeless

by

I will attempt one myself. I will gladly accept a better answer from Travis Brown or Miles Sabin.

Nat can currently not be used to represent large numbers

In the current implementation of Nat, the value corresponds to the number of nested shapeless.Succ[] types:

scala> Nat(3)
res10: shapeless.Succ[shapeless.Succ[shapeless.Succ[shapeless._0]]] = Succ()

So to represent the number 1000000, you would have a type that is nested 1000000 levels deep, which would definitely blow up the scala compiler. The current limit seems to be about 400 from experimentation, but for reasonable compile times it would probably be best to stay below 50.

However, there is a way to encode large integers or other values at type level, provided that you do not want to do calculations on them. The only thing you can do with those as far as I know is to check if they are equal or not. See below.

scala> type OneMillion = Witness.`1000000`.T
defined type alias OneMillion

scala> type AlsoOneMillion = Witness.`1000000`.T
defined type alias AlsoOneMillion

scala> type OneMillionAndOne = Witness.`1000001`.T
defined type alias OneMillionAndOne

scala> implicitly[OneMillion =:= AlsoOneMillion]
res0: =:=[OneMillion,AlsoOneMillion] = <function1>

scala> implicitly[OneMillion =:= OneMillionAndOne]
<console>:16: error: Cannot prove that OneMillion =:= OneMillionAndOne.
       implicitly[OneMillion =:= OneMillionAndOne]
                 ^

This could be used to e.g. enforce same array size when doing bit operations on Array[Byte].

More Related Contents:

  1. Weird behavior trying to convert case classes to heterogeneous lists recursively with Shapeless
  2. Why is the Aux technique required for type-level computations?
  3. Any reason why scala does not explicitly support dependent types?
  4. Use functional combinators on Scala Tuples?
  5. Are HLists nothing more than a convoluted way of writing tuples?
  6. Can’t prove that singleton types are singleton types while generating type class instance
  7. Converting nested case classes to nested Maps using Shapeless
  8. Unable to map on HList
  9. Converting Map[String,Any] to a case class using Shapeless
  10. Understanding the Aux pattern in Scala Type System
  11. Getting subclasses of a sealed trait
  12. What is the purpose of type ascriptions in Scala?
  13. Scala: Why mapValues produces a view and is there any stable alternatives?
  14. Scala Map implementation keeping entries in insertion order?
  15. Modify collection inside a Spark RDD foreach
  16. What is *so* wrong with case class inheritance?
  17. Scala: Why does Seq.contains take an Any argument, instead of an argument of the sequence type?
  18. Customize SparkContext using sparkConf.set(..) when using spark-shell
  19. How to apply a function to a tuple?
  20. What are the differences between asInstanceOf[T] and (o: T) in Scala?
  21. Spark: Transpose DataFrame Without Aggregating
  22. Advantages of F-bounded polymorphism over typeclass for return-current-type problem
  23. Scala Sets contain the same elements, but sameElements() returns false
  24. Can multi-projects from GIT be used as SBT dependencies?
  25. using variable length argument in scala
  26. How to reference external sbt project from another sbt project?
  27. Missing parameter type error by calling toSet?
  28. Dynamically compiling scala class files at runtime in Scala 2.11
  29. How to vectorize DataFrame columns for ML algorithms?
  30. How to instantiate an instance of type represented by type parameter in Scala

Leave a Comment