Conditionals in

About Conditionals

Comparison operators

R has a set of operators that is the same as many other languages:

1 == 2 # FALSE
1 != 2 # TRUE
1 < 2 # TRUE
1 > 2 # FALSE
1 <= 2 # TRUE
1 >= 2 # FALSE

Branching

The basic form of an if-then-else is:

if (x > 10) {
    y <- "big"
} else if (x > 2) {
    y <- "medium"
} else {
    y <- "tiny"
}

In this case, parentheses ( ) around the boolean and braces { } around the body are both required.

A simple if statement can be shortened:

if (x == 0) return("success")

Using vectors in conditionals

Looking ahead to the vector-filtering concept: the conditional clause in parentheses () must evaluate to a single TRUE/FALSE.

Including vectors in the comparison usually gives a vector of booleans, so these will need to be wrapped in an aggregating function such as all() (are all values TRUE?) or any() (is at least one value TRUE).

numbers <- c(4,5,7,9,10,11)
numbers %% 3 == 0 # gives vector of booleans
#> [1] FALSE FALSE FALSE  TRUE FALSE FALSE

# use any() to get a single TRUE/FALSE
if (any(numbers %% 3 == 0)) print("1 or more numbers are divisible by 3")
#> [1] "1 or more numbers are divisible by 3"

The ifelse function

An alternative if-else form may be useful, provided there is only a true/false outcome.

y <- ifelse(x > 8, "big", "small")

# Don't mix result types:
z <- ifelse(x > 100, 100, "small") # numeric or string result? Expect problems!

ifelse() takes exactly three parameters:

1. boolean clause
2. value if TRUE
3. value if FALSE.

Vector inputs are acceptable, and then a vector output is produced. In this case, ensure both branches result in the same type of data: numeric, character, logical, etc. This will be covered in the vector-functions concept.

The two results ("big", "small" in the above example) should be single values. Unfortunately, giving a longer vector here will probably cause R to do something confusing and unexpected, without giving a warning message.

The if_else function

Background

The R language maintainers have a policy of avoiding breaking changes as far as possible. This helps maintain the vast number of third-party packages available in CRAN, but limits the ability to update base R with modern programming concepts.

In recent years, several packages have been released which extend and sometimes replace R functionality, based on multi-decade experience of the strengths and weaknesses of the original R design.

These packages form the tidyverse collection. They were added to the Exercism test runner in 2023, and are available for you to use in exercises.

Our R track predates addition of the tidyverse by several years, so you may notice that many community solutions stick to the base language.

The dplyr package

dplyr is mostly about manipulating dataframes (to be covered later in the course).

Add it with either library(dplyr) at the top of your code, or library(tidyverse) to add the whole collection. You may see warning messages about masked names, but this is rarely a problem.

For the present Concept, the if_else function is worth mentioning. It is mostly similar to ifelse in base R.

library(dplyr) # warnings not shown

x <- 10
if_else(x > 8, "big", "small")
#> [1] "big"

The main difference is that if_else can take an optional fourth argument, for missing values that are represented in R by NA.

xs <- c(3, NA, 10, NA, 4)  # a 5-element vector
xs
#> [1]  3 NA 10 NA  4
if_else(xs > 8, "big", "small", "missing")
#> [1] "small"   "missing" "big"     "missing" "small" 

# contrast:
ifelse(xs > 8, "big", "small")
#> [1] "small" NA      "big"   NA      "small"

You may also find that if_else has much better error messages than the older ifelse.

After loading the dplyr library, you will also have access to the between() function.

R does not allow chaining of comparisons, so 1 <= x <= 5 will give an error message. With dplyr loaded, between(x, 1, 5) is equivalent to x >= 1 & x <= 5, and perhaps more readable. Both lower and upper limits are inclusive.