CODEWITHSUNDEEP
rcombinationstidyr
Jeff Henderson
Jeff Henderson

Reputation: 683

Pulling unique combinations instead of permutations

I'm using the tidyr crossing() function and am able to build a df with each possible permutation of the same names but wish to retain only unique combinations instead of the permutations it seems to create:

names_obj <- c("n1", "n2","n3", "n4","n5","n6","n7","n8","n9","n10","n11","n12")
                    
crossing(n1 = names_obj, n2 = names_obj) %>% 
  filter(n1 != n2)

If you run this code you'll see that in row 1 (n1 = n1, n2 = n10) is the same combo of names as row 12 (n1 = n10, n2 = n1)

How might I adjust this to only keep one of these combos? I don't care which one it is.

Upvotes: 0

Views: 71

Answers (1)

PaulS
PaulS

Reputation: 25323

Another possible solution, however not as succinct as @onyambu's one:

library(tidyverse)

crossing(n1 = names_obj, n2 = names_obj) %>% 
  filter(n1 != n2) %>% 
  filter(!duplicated(str_c(pmax(n1, n2), pmin(n1, n2))))

#> # A tibble: 66 × 2
#>    n1    n2   
#>    <chr> <chr>
#>  1 n1    n10  
#>  2 n1    n11  
#>  3 n1    n12  
#>  4 n1    n2   
#>  5 n1    n3   
#>  6 n1    n4   
#>  7 n1    n5   
#>  8 n1    n6   
#>  9 n1    n7   
#> 10 n1    n8   
#> # … with 56 more rows

Upvotes: 1

Related Questions