R - mutate columns loops over mutated columns when specifying multiple columns to add-CodePudding

following my

Thanks to @julian answer, I can manage this by:

sheet <- sheet %>%
        group_by(across(all_of(.GlobalEnv$filter_list[[select_filter]]))) %>%
        summarise(
          across(
            where(is.numeric),
            list(sum = sum, average = mean, sd = sd),
            .names = "{.col}_{.fn}"
          )
          )

Although this brings out another problem, where if I define my own function, the data is calculated on each value instead of by grouping:

testfunc <- function(.) {
  .   1000
}
sheet <- sheet %>%
        group_by(across(all_of(.GlobalEnv$filter_list[[select_filter]]))) %>%
        summarise(
          across(
            where(is.numeric),
            list(sum = sum, average = mean, sd = sd, test = testfunc),
            .names = "{.col}_{.fn}"
          #total = sum(.data[[names(sheet[,10])]], na.rm = TRUE)
          )
          )

CodePudding user response：

Use a list of functions inside the mutate

sheet <- sheet %>%
  group_by(across(all_of(.GlobalEnv$filter_list[[select_filter]]))) %>%
  mutate(
    across(where(is.numeric), list(sum = sum, average = mean, sd = sd) , .names="{.col}_{.fn}")
  )

But it's better to use summarise instead of mutate

sheet <- sheet %>%
  group_by(across(all_of(.GlobalEnv$filter_list[[select_filter]]))) %>%
  summarise(
    across(where(is.numeric), list(sum = sum, average = mean, sd = sd) , .names="{.col}_{.fn}")
  )

Example with theiris dataset

iris %>%
  mutate(
    across(where(is.numeric), list(sum = sum, average = mean, sd = sd) , .names="{.col}_{.fn}"))