Search code examples
rgroup-bydplyrsummarize

Dynamic groupby function with brackets

df1 <- mtcars %>%
    group_by(gear) %>%
    summarise(Mittelwert = mean(mpg, na.rm = TRUE))
df1

df2 <- mtcars %>%
    group_by(mtcars[[10]]) %>%
    summarise(Mittelwert = mean(mtcars[[1]]), na.rm = TRUE)
df2

The last code gives me the mean of the whole data.frame. Since this code is used in a loop, i need to use brackets. Can you help me to get a dynamic code with valid results?

Solution

  • We can use group_by_at and summarise_at to specify column number if we want to avoid using names.

    library(dplyr)
mtcars %>%
  group_by_at(10) %>%
  summarise_at(1, mean, na.rm = TRUE)

# A tibble: 3 x 2
#   gear   mpg
#  <dbl> <dbl>
#1  3.00  16.1
#2  4.00  24.5
#3  5.00  21.4

    which is equivalent to 

    mtcars %>%
    group_by(gear) %>%
    summarise(Mittelwert = mean(mpg, na.rm = TRUE))

#   gear Mittelwert
#  <dbl>      <dbl>
#1  3.00       16.1
#2  4.00       24.5
#3  5.00       21.4