Correct syntax for mutate_if

I learned this trick from the purrr tutorial, and it also works in dplyr. There are two ways to solve this problem:
First, define custom functions outside the pipe, and use it in mutate_if():

any_column_NA <- function(x){
replace_NA_0 <- function(x){
mtcars %>% mutate_if(any_column_NA,replace_NA_0)

Second, use the combination of ~,. or .x.( .x can be replaced with ., but not any other character or symbol):

mtcars %>% mutate_if(~ any(,~ if_else(,0,.x))
#This also works
mtcars %>% mutate_if(~ any(,~ if_else(,0,.))

In your case, you can also use mutate_all():

mtcars %>% mutate_all(~ if_else(,0,.x))

Using ~, we can define an anonymous function, while .x or . stands for the variable. In mutate_if() case, . or .x is each column.

Leave a Comment