Checking arguments

Introduction to Writing Functions in R

Richie Cotton

Data Evangelist at DataCamp

The geometric mean

calc_geometric_mean <- function(x, na.rm = FALSE) {
  x %>%
    log() %>%
    mean(na.rm = na.rm) %>%
    exp()
}

calc_geometric_mean(letters)

Error in log(.) : non-numeric argument to mathematical function

Checking for numeric values

calc_geometric_mean <- function(x, na.rm = FALSE) {

  if(!is.numeric(x)) {

    stop("x is not of class 'numeric'; it has class '", class(x), "'.")

  }

  x %>%
    log() %>%
    mean(na.rm = na.rm) %>%
    exp()
}

Error in calc_geometric_mean(letters) : 
  x is not of class 'numeric'; it has class 'character'.

assertive makes errors easy

Checking types of inputs

assert_is_numeric()
assert_is_character()
is_data.frame()
...
is_two_sided_formula()
is_tskernel()

Using assertive to check x

calc_geometric_mean <- function(x, na.rm = FALSE) {

  assert_is_numeric(x)

  x %>%
    log() %>%
    mean(na.rm = na.rm) %>%
    exp()
}

 Error in calc_geometric_mean(letters) : 
  is_numeric : x is not of class 'numeric'; it has class 'character'.

Checking x is positive

calc_geometric_mean <- function(x, na.rm = FALSE) {
  assert_is_numeric(x)

  assert_all_are_positive(x)

  x %>%
    log() %>%
    mean(na.rm = na.rm) %>%
    exp()
}

calc_geometric_mean(c(1, -1))

Error in calc_geometric_mean(c(1, -1)) :
  is_positive : x contains non-positive values.
There was 1 failure:
  Position Value   Cause
1        2    -1 too low

is_* functions

assert_is_numeric()
assert_all_are_positive()

is_numeric() (returns logical value)
is_positive() (returns logical vector)
is_non_positive()

Custom checks

calc_geometric_mean <- function(x, na.rm = FALSE) {
  assert_is_numeric(x)

  if(any(is_non_positive(x), na.rm = TRUE)) {
    stop("x contains non-positive values, so the geometric mean makes no sense.")
  }

  x %>%
    log() %>%
    mean(na.rm = na.rm) %>%
    exp()
}

calc_geometric_mean(c(1, -1))

Error in calc_geometric_mean(c(1, -1)) : 
  x contains non-positive values, so the geometric mean makes no sense.

Fixing input

use_first(c(1, 4, 9, 16))

[1] 1
Warning message:
Only the first value of c(1, 4, 9, 16) (= 1) will be used.

coerce_to(c(1, 4, 9, 16), "character")

[1] "1"  "4"  "9"  "16"
Warning message:
Coercing c(1, 4, 9, 16) to class ‘character’.

Fixing na.rm

calc_geometric_mean <- function(x, na.rm = FALSE) {
  assert_is_numeric(x)
  if(any(is_non_positive(x), na.rm = TRUE)) {
    stop("x contains non-positive values, so the geometric mean makes no sense.")
  }

  na.rm <- coerce_to(use_first(na.rm), target_class = "logical")

  x %>%
    log() %>%
    mean(na.rm = na.rm) %>%
    exp()
}

calc_geometric_mean(1:5, na.rm = 1:5)

[1] 2.605171
Warning messages:
1: Only the first value of na.rm (= 1) will be used. 
2: Coercing use_first(na.rm) to class ‘logical’.

Let's practice!

Introduction to Writing Functions in R