CODEWITHSUNDEEP
rgroup-byshinydplyr
Neil
Neil

Reputation: 8247

dplyr - groupby on multiple columns using variable names

I am working with R Shiny for some exploratory data analysis. I have two checkbox inputs that contain only the user-selected options. The first checkbox input contains only the categorical variables; the second checkbox contains only numeric variables. Next, I apply a groupby on these two selections:

var1 <- input$variable1      # Checkbox with categorical variables
var2 <- input$variable2      # Checkbox with numerical variables

v$data <- dataset %>%
  group_by_(var1) %>%
  summarize_(Sum = interp(~sum(x), x = as.name(var2))) %>%
  arrange(desc(Sum))

When only one categorical variable is selected, this groupby works perfectly. When multiple categorical variables are chosen, this groupby returns an array with column names. How do I pass this array of column names to dplyr's groupby?

Upvotes: 49

Views: 61779

Answers (4)

MrFlick
MrFlick

Reputation: 206177

dplyr version >1.0

With more recent versions of dplyr, you should use across along with a tidyselect helper function. See help("language", "tidyselect") for a list of all the helper functions. In this case if you want all columns in a character vector, use all_of()

cols <- c("mpg","hp","wt")
mtcars %>% 
   group_by(across(all_of(cols))) %>% 
   summarize(x=mean(gear))

original answer (older versions of dplyr)

If you have a vector of variable names, you should pass them to the .dots= parameter of group_by_. For example:

mtcars %>% 
   group_by_(.dots=c("mpg","hp","wt")) %>% 
   summarize(x=mean(gear))

Upvotes: 66

bschneidr
bschneidr

Reputation: 6278

You can use the helpers from rlang package, which is created by the same team that created dplyr. When using dplyr and other tidyverse packages, you don't have to load the rlang packages in order to use those helpers.

Specifically, you can use the syms function and the !!! function like so:

library(dplyr)

group_cols <- c("vs", "am")

mtcars %>% 
  group_by(!!!syms(group_cols)) %>% 
  summarize(mean_wt = mean(wt))

This closely-related question and answer explains how the !! operator and sym function are used for a single column name (i.e. a length-one character vector).

Upvotes: 10

Michael M
Michael M

Reputation: 1593

With dplyr 1.0.0, we have the following possibility based on the "normal" group_by:

library(dplyr)

group_cols <- c("vs", "am")

mtcars %>% 
  group_by(across(all_of(group_cols))) %>% 
  summarize(mean_wt = mean(wt))

Upvotes: 11

bschneidr
bschneidr

Reputation: 6278

Recent versions of the dplyr package include variants of group_by, such as group_by_if and group_by_at. You can use these to perform column selections with syntax that is similar to the select function.

Just as you could select a list of columns with select(my_data, one_of(group_cols)), you can use group_by_at to do the following:

library(dplyr)

group_cols <- c("vs", "am")

mtcars %>% 
  group_by_at(.vars = vars(one_of(group_cols))) %>% 
  summarize(mean_wt = mean(wt))

Upvotes: 5

Related Questions