CODEWITHSUNDEEP
rfunctiondplyr
J.K.
J.K.

Reputation: 381

Taking ... (three dots) argument for grouping variables in dplyr and use ... as names of a new data frame in a function

Purpose

I would like to take ... (three dots) argument for grouping variables in dplyr and use ... as names of a new data frame in a function. Question section includes details about what I want to achieve.

Sample Data

library(tidyverse)
  library(tibble)
  library(data.table)
  
  rename <- dplyr::rename
  select <- dplyr::select

  set.seed(10002)
  id <- sample(1:20, 1000, replace=T)
  
  set.seed(10003)
  group1 <- sample(letters, 1000, replace=T)
  
  set.seed(10004)
  group2 <- sample(LETTERS, 1000, replace=T)

  
  df <-
    data.frame(id, group1, group2)

Question

fn <- function(df, ...){
    
    group_ <- enquos(...)
    
# First I will use this as grouping variables in dplyr
    df %>%
      group_by(!!!group_) %>% 
      summarise(obs = n()) 

# The question is the second operation.     
      # I would like to create a data frame with NAs here so that I can rbind using for loop later
      # for example, if ... = group1 
      # f <- data.frame(id = NA, group1 = NA, output = NA)
      # for example, if ... = group1, group2
      # f <- data.frame(id = NA, group1 = NA, group1 = NA, output = NA)
      # Is there a way to take the ... argument abd use them as column names in a new data frame f in a function? 
      



  }

Upvotes: 1

Views: 471

Answers (1)

akrun
akrun

Reputation: 887213

After we create the grouping attributes, get the group column names directly with group_vars, then create the dataset dynamically using those names

fn <- function(df, ...){
    
    group_ <- enquos(...)
    

  
  tmp <-   df %>%
      group_by(!!!group_) %>% 
      summarise(obs = n(), .groups = 'keep')

  nm1 <- group_vars(tmp) 
  tibble::as_tibble(setNames(rep(list(NA), length(nm1) + 2),
        c('id', nm1, 'output')))

}

-testing

> fn(df, group1)
  # A tibble: 1 x 3
  id    group1 output
  <lgl> <lgl>  <lgl> 
1 NA    NA     NA    
> fn(df, group1, group2)
  # A tibble: 1 x 4
  id    group1 group2 output
  <lgl> <lgl>  <lgl>  <lgl> 
1 NA    NA     NA     NA

Upvotes: 1

Related Questions