Using a custom function with dplyr summarise

April 15, 2023

I have a data frame which I apply some custom summarization logic to (if maximum value per group is over 100, use the maximum value, otherwise use the average of values per group):

library(tidyverse)

dat <- data.frame(grp = c("a", "a", "a", "b", "b"),
           vals = c(115, 100, 101, 90, 100))

dat %>% 
  group_by(grp) %>% 
  summarise(new_val = case_when(max(vals) >= 100 ~ max(vals),
                                TRUE ~ mean(vals)))

However, I need to reuse the custom summarization logic with different maximum cutoff values but don’t want to hardcode in the summarization logic over and over. How can I create a function to do so?

The following does not work:

sumFunc <- function(max_val) {
  case_when(max(vals) >= max_val ~ max(vals),
            TRUE ~ mean(vals))
}

dat %>% 
  group_by(grp) %>% 
  summarise(new_val = sumFunc(100))

>Solution :

Or this:

sumFunc <- function(col,max_val) {
  case_when(max(col) >= max_val ~ max(col),
            TRUE ~ mean(col))
}

dat %>% 
  group_by(grp) %>% 
  summarise(new_val = sumFunc(vals,100))

  grp   new_val
  <fct>   <dbl>
1 a         115
2 b          95