R - All pairwise combinations of column strings concatenated by the row

问题

How can I transform data frames like this one:

  X__1  X__2  X__3 
  <chr> <chr> <chr>
1 a     b     c    
2 d     e     f    
3 g     h     i    
4 j     k     l

Into this one:

  X__1  X__2  X__3 
  <chr> <chr> <chr>
1 a-d   b-e   c-f  
2 a-g   b-h   c-i  
3 a-j   b-k   c-l  
4 d-g   e-h   f-i  
5 d-j   e-k   f-l  
6 g-j   h-k   i-l

In other words, it should make all possible pairwise combinations of the whole rows in the data frame, combining strings from the same column but separated by a sign(-). It does not need to repeat an already made combination in the other order of mentioning the letter, i.e. "a-d, b-e, c-f" is required but not "d-a, e-b, f-c".

Thank you in advance. Let me know how to improve posing the question if needed.

回答1:

We can use map

library(purrr)
library(stringr)
map_dfc(df1, combn, m = 2, FUN = str_c, collapse="-")
# A tibble: 6 x 3
#  X__1  X__2  X__3 
#  <chr> <chr> <chr>
#1 a-d   b-e   c-f  
#2 a-g   b-h   c-i  
#3 a-j   b-k   c-l  
#4 d-g   e-h   f-i  
#5 d-j   e-k   f-l  
#6 g-j   h-k   i-l

Or using summarise/unnest

library(dplyr)
library(tidyr)
df1 %>%
    summarise(across(everything(), ~ 
        list(combn(., 2, FUN = str_c, collapse="-")))) %>%
    unnest(everything())
# A tibble: 6 x 3
#  X__1  X__2  X__3 
#  <chr> <chr> <chr>
#1 a-d   b-e   c-f  
#2 a-g   b-h   c-i  
#3 a-j   b-k   c-l  
#4 d-g   e-h   f-i  
#5 d-j   e-k   f-l  
#6 g-j   h-k   i-l

Or with base R

data.frame(lapply(df1, combn, m = 2, paste, collapse="-"))
#   X__1 X__2 X__3
#1  a-d  b-e  c-f
#2  a-g  b-h  c-i
#3  a-j  b-k  c-l
#4  d-g  e-h  f-i
#5  d-j  e-k  f-l
#6  g-j  h-k  i-l

data

df1 <- structure(list(X__1 = c("a", "d", "g", "j"), X__2 = c("b", "e", 
"h", "k"), X__3 = c("c", "f", "i", "l")), class = "data.frame", row.names = c("1", 
"2", "3", "4"))

来源：https://stackoverflow.com/questions/62251966/r-all-pairwise-combinations-of-column-strings-concatenated-by-the-row

标签

combinations

rows