Subsetting columns from a dataframe to create new dataframes by iteration

Question

I have a large dataframe and want to create several new dataframes from individual columns. Each new dataframe should consist of the first two columns plus one of the next columns.

I can do this manually; my desired dummy products being subA, subB, and subC.

library(tidyverse)

# dummy dataframe
chr1 <- letters[1:4]
chr2 <- letters[23:26]
A <- 1:4
B <- 5:8
C <- 9:12
data <- tibble(chr1, chr2, A, B, C)

# manual subsetting
subA <- data %>%
  select(chr1, chr2, A)

subB <- data %>%
  select(chr1, chr2, B)

subC <- data %>%
  select(chr1, chr2, C)

However, my original dataframe has about 700 columns, so I am looking for some kind of simple iteration, preferably in tidyverse (purrr), that is flexible enough to be supplemented by pipes into additional functions.

Answer 1

in base:

lapply(split.default(data[-c(1,2)], names(data[-c(1,2)])), function(x) 
        cbind(data[c(1,2)], x))

or in tidyverse:

data %>% 
  pivot_longer(-c(1, 2)) %>% 
  split(., .$name) %>% 
  map(., ~pivot_wider(.x))

#> $A
#> # A tibble: 4 x 3
#>   chr1  chr2      A
#>   <chr> <chr> <int>
#> 1 a     w         1
#> 2 b     x         2
#> 3 c     y         3
#> 4 d     z         4
#> 
#> $B
#> # A tibble: 4 x 3
#>   chr1  chr2      B
#>   <chr> <chr> <int>
#> 1 a     w         5
#> 2 b     x         6
#> 3 c     y         7
#> 4 d     z         8
#> 
#> $C
#> # A tibble: 4 x 3
#>   chr1  chr2      C
#>   <chr> <chr> <int>
#> 1 a     w         9
#> 2 b     x        10
#> 3 c     y        11
#> 4 d     z        12

If you want to save these as individual dataframes:

list2env(
    lapply(split.default(data[-c(1,2)], paste0("sub", names(data[-c(1,2)]))), function(x) 
           cbind(data[c(1,2)], x)),  
    envir=.GlobalEnv)

^{Created on 2024-01-16 with reprex v2.0.2}