I have a problem with add_xy_position() after using tukey_hsd() from the rstatix package inside a function. This is what my function looks like:

make_tukey_test <- function (data,variable,grouping_variable){
  data %>% 
    tukey_hsd(variable ~ grouping_variable)
}

When I call the function using the following code, it works perfectly fine and the test results are being saved in the data.frame:

test <- make_tukey_test(data = dat, variable = dat$num_var, grouping_variable = dat$factor_var)

However, when I try to add the x and y coordinates of dat$factor_var using add_xy_position() like this:

test <- test %>%
  add_xy_position(x = “factor_var”)

I get the following error message:

Error: Must group by variables found in `.data`. * Column `grouping_variable` is not found.

However, when I use tukey_hsd() outside of my function, the code works perfectly fine and the coordinates are being added to the data.frame.

I would greatly appreciate some helpful suggestions, as I have no clue why the code doesn't work when I´m using it inside my function.

CodePudding user response：

I would suggest to pass column names as strings. Since tukey_hsd accepts a formula object you can use reformulate to create that and pass the grouping_variable as it is in add_xy_position function.

Here is an example with mtcars dataset.

library(rstatix)
library(dplyr)

dat <- mtcars %>% mutate(cyl = factor(cyl)) 

make_tukey_test <- function (data,variable,grouping_variable){
  data %>% 
    tukey_hsd(reformulate(grouping_variable, variable)) %>%
    add_xy_position(x = grouping_variable)
}

test <- make_tukey_test(data = dat, variable = "mpg", grouping_variable = "cyl")