dplyr summarise_each() using multiple functions for different column subsets across the same groups

Question

I'd like to use summarise_each() to apply multiple functions to a grouped dataset. However, rather than apply each function to all columns, I'd like to apply each function to particular subsets. I realize I could do this by specifying each column with summarise(), but I have many variables.

Is there an alternate solution to either 1) using summarise_each() and then deleting the unneeded columns or 2) saving the group_by() result, performing multiple separate summarise_each() operations and combining the results?

If this is not clear, let me know and I can try to illustrate with some example code.

Answer 1

I would suggest the following: here I would like to apply min function to one variable and max function to other. Then I simply merge those with the grouping variable.

> by_species <- iris %>% group_by(Species)    

Start with variable for which I want to apply the min function:

min_var <- by_species %>% summarise_each(funs(min), Petal.Width) min_var Source: local data frame [3 x 2]

      Species Petal.Width
       (fctr)       (dbl)
1     setosa         0.1
2 versicolor         1.0
3  virginica         1.4

Then the variable for which I want to apply the max function:

max_var <- by_species %>% summarise_each(funs(max), Sepal.Width) max_var Source: local data frame [3 x 2]

     Species Sepal.Width
      (fctr)       (dbl)
 1     setosa         4.4
 2 versicolor         3.4
 3  virginica         3.8

Now, we just merge the above two:

left_join(min_var,max_var) Joining by: "Species" Source: local data frame [3 x 3]

      Species Petal.Width Sepal.Width
     (fctr)       (dbl)       (dbl)
1     setosa         0.1         4.4
2 versicolor         1.0         3.4
3  virginica         1.4         3.8