Question

Given

a = [[:a, :b, :c]]

1) I understand this

a.each{|(x, y), z| p z} # => :b

that there are two variables (x, y) and z, so the third element :c is thrown away, and z matches :b. And I understand this

a.each{|(x, y), z| p y} # => nil

that (x, y) matches :a, and since it is not an array, there are no elements to it, and so y matches nil.

But how does

a.each{|(x, y), z| p x} # => :a

work? I expect nil to be returned.

2) Why are the return values like this?

a.each{|(x, y)| p x} #=> :a
a.each{|(x, y)| p y} #=> :b

I expect them to both return nil.

Answer 1

It’s because of the syntax of parallel assignment.

a = [[:a, :b, :c]]

So a.each has only one element to iterate, which is [:a, :b, :c].

In the first case:

(x, y), z = [:a, :b, :c]
#=> x == :a, y == nil, z == :b

Here (x, y) is an array to match the first element :a, and x gets it, then z simply matches the second element :b.

And in the second case:

(x, y) = [:a, :b, :c]
#=> x == :a, y == :b

Here (x, y) as an entire array matches the array [:a, :b, :c], so x and y get :a and :b respectively.

This is just like requiring the “args + optional args (keyword args) + rest args” combination match provided arguments. It is just smart enough to take arguments by sequence.

Another smart example:

(a,b) = 1,2
=> [1, 2] # array match
#=> a == 1, b == 2

(a,b)=[1,2]
=> [1, 2] # array match
#=> a == 1, b == 2

In either case above, it will simply make the best guess on what it should take.