In R, how do I get an if statement to recognize if input in double curly brackets is a certain value?

user_9877445 0
Viewed 49

I'm building a function where users can select a column from a list of options, and within the function, I want to do an if statement where, if the x variable is one of the options that the function is designed to work with, it will generate the plot. Otherwise, it will print an error message.

However, when I try to do if({{x_variable}} == ses)--the correct variable in this example--I keep getting

Error in make_plot(test_data, x_variable = ses) : object 'ses' not found

That's the same error I get for if(enquo(x_variable) == ses) and if(!!x_variable == ses)

The correct answer will produce the plot when x_variable is ses and will print the error when x_variable is anything else.

Here's a sample dataset and my function (that does not work collectively, but each individual part does):

library(dplyr)
library(rlang) #if enquo() is needed

test_data <- tibble(ses = c(rep(c("High", "Mid", "Mid Low", "Low"), 2)),
                    total = c(10, 20, 20, 30, 9, 11, 40, 60))

make_plot <- function(data, x_variable) {

  if({{x_variable}} == "ses") {
ggplot(data = test_data, aes(x = {{x_variable}}, y = total)) +
  geom_col()
  } else {
    print("This function isn't designed for this variable, sorry!")
  }
}

make_plot(test_data, x_variable = ses)
make_plot(test_data, x_variable = anything_else)
rlang curly-braces r
2 Answers

You have to use:

make_plot(test_data, x_variable = "ses")

or alternatively:

ses <- "ses"
make_plot(test_data, x_variable = ses)

This error means that the object ses is not declared.

If you want to be able to pass undeclared objects such as ses as an input, you could use substitute(x_variable) or deparse(substitute(x_variable))

make_plot <- function(data, x_variable) {
  #print(deparse(substitute(x_variable)))
  if(deparse(substitute(x_variable)) == "ses") {
    ggplot(data = test_data, aes(x = ses, y = total)) +
      geom_col()
  } else {
    print("This function isn't designed for this variable, sorry!")
  }
}

enter image description here

This is non-standard evaluation however, so make sure this is indeed what you're after as it can lead to surprising behaviours.

This explains the difference between both options, from Advanced R enter image description here

user_52242360

To supplement @gaut's elaborate answer, or if it's easier for you to remember (or understand how {{ work), you would need the following to use the curly brackets.

names(select(data, {{x_variable}})) == "ses"

user_47864660
Related