Search code examples
scalalanguage-design

Why use curly braces over parentheses?


In a lot of Scala examples I see people use curly braces in places I find outright strange, when the same statement could easily be written using parentheses.

Example:

lst foreach (x => println(s"the value returned is: $x")) // parens
lst foreach {x => println(s"you get the idea, $x")} // braces

I understand that you can use braces as an alternative to parentheses, simply because it allows you to write a statement on multiple lines:

val res = for {
  x <- coll1
  y <- coll2
} yield (x, y) 
  • So when it's written on a single line, is there any inherent reason to use one over the other?
  • The outcome should be the same in the end, or am I missing something?
  • Or is it simply just a matter of style and/or personal taste?

Solution

  • In general, there are many cases when you would prefer curly braces (e.g. multiline expressions, for comprehensions), but let's talk specifically about

    when it's written on a single line, is there any inherent reason to use one over the other

    In a second case it's not just curly braces, instead of parentheses, it's curly braces with ommited parentheses. Scala allows you to ommit parenthesis sometimes, and the later syntax is used to access to the niceties you got in partial functions (namely, pattern matching), so

    lst foreach {x => println(s"you get the idea, $x")}
    

    is actually

    lst foreach({x => println(s"you get the idea, $x")})
    

    which, as I said, can be useful from pattern matching POV:

    val map = Map("foo" -> "bar")
    map foreach { case (k, v) => println(s"key: $k, value: $v") }
    // which is not possible with the usual parenthesis