Programming Scala, 3rd Edition

Chapter 20. Dynamic Invocation in Scala

Most of the time, Scala’s static typing is a virtue. It adds safety constraints that are useful for ensuring correctness at runtime and easier comprehension when browsing code. Many errors are caught at compile time. These benefits are especially useful in large-scale systems.

Occasionally, you might miss the benefits of dynamic typing, however. For example, in “Structural Types”, we discussed a scenario where we would like to process SQL query result sets without necessarily defining a custom type, like a case class, for each and every query’s returned record type. We’ll explore this scenario in this chapter. We’ll also avoid the completely untyped alternative of holding the column names and values as key-value pairs in a map or using a generic record type that requires casting column values to the correct type.

We’ll leverage two mechanisms in Scala 3 that fall in the gap between specific types and generic records. The first is the new Scala 3 feature called structural types that I introduced in “Structural Types”. This chapter will start with a more sophisticated example for the query scenario. Then we’ll discuss a less type-safe but more flexible mechanism that exists in both Scala 2 and 3, the scala.Dynamic trait. The Dynamic trait provides greater flexibility, but structural types provide better type safety.

Structural Types Revisited

Scala 3 structural types provide type-safe dynamic invocation. We saw a short example that uses the ...