CODEWITHSUNDEEP
rdplyr
lukitamodric
lukitamodric

Reputation: 37

join tibbles by unordered list

Suppose I have two tibbles: a and b.

a <- tibble("unordered_characters" = list( c("A","B"), c("X","Y") ) ,
            "v" = c("G", "F"))
b <- tibble("unordered_characters" = list( c("B","A"), c("Y","X")),
            "x" = c("M", "d"))

I want to join b on a using their common variable, which is a list. But I want the list to be unordered, so that c("A","B") should be the same as c("B","A"). In other words, setequal(c("A","B"),c("B","A")) returns TRUE. This would to create the following table:

unordered_characters v x
------------------------
c("A","B")            G M 
c("X","Y")           F d

But a %>% left_join(b) results in the following:

unordered_characters v x
------------------------
c("A","B")            G NA 
c("X","Y")           F NA

How can I fix this?

Upvotes: 2

Views: 38

Answers (1)

akrun
akrun

Reputation: 887118

We could use map to loop over the list and sort before doing the join

library(dplyr)
library(purrr)
a %>% 
  mutate(unordered_characters = map(unordered_characters, sort)) %>% 
  left_join(b %>% 
       mutate(unordered_characters = map(unordered_characters, sort)))
# A tibble: 2 x 3
#  unordered_characters v     x    
#  <list>               <chr> <chr>
#1 <chr [2]>            G     M    
#2 <chr [2]>            F     d

Upvotes: 1

Related Questions