Here are two functions:

f <- function(a,b) {
  print(a^2)
  print(b^2)
}

y <- function(a,b) {
  a^2
  b^2
}

Question: Why do I get [1] 4 [1] 16 with f(2, 4) but get only [1] 16 with y(2, 4), which completely ignoring one of my arguments?

---

First, keep in mind that:

1. Every line in a function is evaluated.
2. The last expression evaluated is (by default) the returned value. Alternatively one can use return to return a specific object from a function.

So let’s take a look at the two functions, starting from the second one, which should be easier according to the rules above:

y <- function(a, b) {
  a^2
  b^2  # This is the last expression
}

As described earlier, the only returned value should be b^2, although a^2 is also evaluated. Hence y(2, 4) returns 16.

Now let’s see the other:

f <- function(a, b) {
  print(a^2)
  print(b^2)  # This is still the last expression
}

print is a function which simply prints out the result of the passed expression as long as it is not causing errors. By just evaluating print, it prints. Thus, you see 4 and 16 printed out when you execute f(2, 4).

However, still only 16 is returned by f. You can try:

result <- f(2, 4)
result  # 16

You’ll see 4 and 16 are printed again, but result is just 16.

If you really need multiple objects to be returned, usually we will use a list (or a named vector in really simple cases like yours):

# In a list
y_list <- function(a, b) {
  list(a_square = a^2, b_square = b^2)
}

# In a named vector
y_vec <- function(a, b) {
  c(a_square = a^2, b_square = b^2)
}

Now you can do y_list(2, 4) or y_vec(2, 4), store the returned value, and subset what you need in the following analyses. The advantage of returning multiple values in a list is that those values don’t have to be in the same class, which is a nice characteristic of lists.