How to get all possible subsets of a character vector in R?

Question

Having the following vector:

c("test1","test2","test3")

I am trying to get a list or data frame containing the following entries:

"test1" "test2" "test3"
"test1" "test2" NA
"test1" NA "test3"
"test1"  NA NA
NA  "test2" "test3"
NA  "test2" NA
NA  NA "test3"

The goal would be to get all possible subsets while the order doesn't matter, that is "text1" "text2" NA is equivalent to "text2" "text1" NA. I very much appreciate any help!

Thanks for the question. It's not important but I included it because if I get a data frame there needs to be an entry indicating that this string is missing. — Patrick Balada, Mar 24 '16 at 09:42
Those aren't all the possible subsets... e.g. `"test1",NA,NA` is missing — digEmAll, Mar 24 '16 at 09:45
@digEmAll This is what they mean by "*and so forth*" I would guess... — David Arenburg, Mar 24 '16 at 09:46
Pretty much the same as the second part of [this](http://stackoverflow.com/questions/27953588/unordered-combinations-in-r/27953641#27953641) — Rich Scriven, Mar 24 '16 at 10:36

score 14 · Accepted Answer · answered Mar 24 '16 at 10:09

You can use combn:

res <- unlist(lapply(1:3, combn, 
                     x = c("test1","test2","test3"), simplify = FALSE), 
              recursive = FALSE)
res <- sapply(res, `length<-`, 3)
#        [,1]    [,2]    [,3]    [,4]    [,5]    [,6]    [,7]   
#[1,] "test1" "test2" "test3" "test1" "test1" "test2" "test1"
#[2,] NA      NA      NA      "test2" "test3" "test3" "test2"
#[3,] NA      NA      NA      NA      NA      NA      "test3"

score 8 · Answer 2 · answered Mar 24 '16 at 10:07

There is a package sets with the relevant function.

library(sets)
a <- c("test1","test2","test3")
set_power(a)

{{}, {"test1"}, {"test2"}, {"test3"}, {"test1", "test2"}, {"test1", "test3"}, {"test2", "test3"}, {"test1", "test2", "test3"}}

This returns the set of all subsets.

zx8754 · Answer 3 · 2016-03-24T10:28:56.933

Using combn, and data.table::rbindlist with fill = TRUE option to make NA values.

#data
a <- c("test1","test2","test3")

#result
data.table::rbindlist(
        sapply(1:3, function(i) as.data.frame(t(combn(a, i)))), fill = TRUE)

#output
#       V1    V2    V3
# 1: test1    NA    NA
# 2: test2    NA    NA
# 3: test3    NA    NA
# 4: test1 test2    NA
# 5: test1 test3    NA
# 6: test2 test3    NA
# 7: test1 test2 test3

How to get all possible subsets of a character vector in R?

3 Answers3

Linked

Related