Is there a way to use mutate inside of ones own function?

user_13734451 0
Viewed 73

Am trying to use mutate function inside my own created function to in order to deal with DRY principle. Am not sure if it would be possible. Whenever I run the function collapsevars I get the correct answer but the mutate function does not pick the variable name specified in the collapsevars function. Any leads? The sample code is as below:

library(dplyr)

set.seed(2021)

avars <- c("a1", "a2") # Variables to collapse to a single variable
bvars <- c("b1", "b2") # Variables to collapse to another variable

mydata <- tibble(
  a1 = factor(round(runif(20, 1, 3)),
              labels = c("Yes", "No", "N/A")),
  a2 = factor(round(runif(20, 1, 3)),
              labels = c("Yes", "No", "N/A")),
  b1 = factor(round(runif(20, 1, 3)),
              labels = c("Yes", "No", "N/A")),
  b2 = factor(round(runif(20, 1, 3)),
              labels = c("Yes", "No", "N/A")),

  c1 = round(rnorm(20, 10, 2)))
mydata

mydata <- mydata %>%
  mutate(afin = ifelse(rowSums(
    select(., all_of(avars)) == "Yes") > 0,
                       "Yes", "No"),
         bfin = ifelse(rowSums(
           select(., all_of(bvars)) == "Yes") > 0,
                       "Yes", "No"))

collapsevars <- function(var, varcollapse){
  mydata <- mydata %>%
    mutate(var = ifelse(rowSums(
      select(., all_of(varcollapse)) == "Yes") > 0,
                         "Yes", "No"))

}

collapsevars(afin, avars)
dplyr r
3 Answers

You can also use the following solution. We use unquo or ensym function here to unquote a user-defined argument and then we use bang-bang !! operator on the LHS of := to force-evaluate it:

library(rlang)

collapsevars <- function(var, varcollapse){
  var <- enquo(var)
  
  mydata <- mydata %>%
    mutate(!!var := ifelse(rowSums(
      select(., all_of(varcollapse)) == "Yes") > 0,
      "Yes", "No"))
  
  mydata
}

collapsevars(afin, avars)

# A tibble: 20 x 7
   a1    a2    b1    b2       c1 afin  bfin 
   <fct> <fct> <fct> <fct> <dbl> <chr> <chr>
 1 No    Yes   Yes   Yes      10 Yes   Yes  
 2 N/A   N/A   N/A   No       14 No    No   
 3 No    No    No    Yes       9 No    Yes  
 4 No    No    N/A   No       14 No    No   
 5 No    No    Yes   N/A      10 No    Yes  
 6 No    N/A   N/A   Yes       8 No    Yes  
 7 No    N/A   N/A   Yes      13 No    Yes  
 8 No    N/A   Yes   No        9 No    Yes  
 9 N/A   N/A   Yes   No       11 No    Yes  
10 N/A   N/A   N/A   Yes      11 No    Yes  
11 Yes   Yes   No    Yes       9 Yes   Yes  
12 N/A   N/A   No    No        5 No    No   
13 No    N/A   Yes   Yes      10 No    Yes  
14 No    N/A   No    N/A       9 No    No   
15 N/A   No    No    Yes       8 No    Yes  
16 No    Yes   Yes   No       10 Yes   Yes  
17 No    N/A   No    No       11 No    No   
18 N/A   Yes   No    No       10 Yes   No   
19 N/A   N/A   N/A   No        7 No    No   
20 No    No    N/A   N/A       7 No    No
user_143145200

Using case_when

library(dplyr)
collapsevars <- function(var, varcollapse){
 mydata %>%
    mutate({{var}} := case_when(rowSums(
          select(., all_of(varcollapse)) == "Yes") > 0 ~ 
      "Yes", TRUE ~ "No"))
 
}
collapsevars(afin, avars)
user_37322710

You can use {{col}} := on left hand side in mutate -

library(dplyr)

collapsevars <- function(var, varcollapse){
  mydata <- mydata %>%
    mutate({{var}} := ifelse(rowSums(
      select(., all_of(varcollapse)) == "Yes") > 0,
      "Yes", "No"))
  mydata
}

collapsevars(afin, avars)

# A tibble: 20 x 6
#    a1    a2    b1    b2       c1 afin 
#   <fct> <fct> <fct> <fct> <dbl> <chr>
# 1 No    Yes   Yes   Yes      10 Yes  
# 2 N/A   N/A   N/A   No       14 No   
# 3 No    No    No    Yes       9 No   
# 4 No    No    N/A   No       14 No   
# 5 No    No    Yes   N/A      10 No   
# 6 No    N/A   N/A   Yes       8 No   
# 7 No    N/A   N/A   Yes      13 No   
# 8 No    N/A   Yes   No        9 No   
# 9 N/A   N/A   Yes   No       11 No   
#10 N/A   N/A   N/A   Yes      11 No   
#11 Yes   Yes   No    Yes       9 Yes  
#12 N/A   N/A   No    No        5 No   
#13 No    N/A   Yes   Yes      10 No   
#14 No    N/A   No    N/A       9 No   
#15 N/A   No    No    Yes       8 No   
#16 No    Yes   Yes   No       10 Yes  
#17 No    N/A   No    No       11 No   
#18 N/A   Yes   No    No       10 Yes  
#19 N/A   N/A   N/A   No        7 No
user_39629140
Related