How can I select columns/rows with the opposite of the column/row names in R?

Question

I have a matrix or a data-frame with specific column names. With a vector containing some of the column names, I can easily address those columns of the matrix. But is there as well a simple way to address the opposite columns, which are not listed in the vector:

mat <- matrix(c(1:12), ncol=4)
colnames(mat) <- c("a", "b", "c", "d")
not_target_col <- c("a", "b")

in this case I like to have the columns c and d. I search something like this, without making the extra step:

pos <- colnames(mat) != not_target_col
mat[,pos]

Additional explanation

I like to make it more clear: When I have a numeric vector I can get the opposite when I add *-1

not_target_col <- c(1,2)
mat[,not_target_col * -1]

There is as well a technique like this when I use a logical vector. Here I just have to add a !.

not_target_col <- c(T,T,F,F)
mat[,!not_target_col]

@akrun definitly, what i do now? – and-bri Mar 20 '17 at 06:48 — and-bri, Mar 20 '17 at 06:48
Nothing you can do. Somebody will dupe tag it – akrun Mar 20 '17 at 06:49 — akrun, Mar 20 '17 at 06:49
I added another solution below – akrun Mar 20 '17 at 06:56 — akrun, Mar 20 '17 at 06:56

score 0 · Accepted Answer · answered Mar 20 '17 at 06:40

We can use setdiff between column names (colnames) and not_target_col to get the column names which do not match with not_target_col.

setdiff(colnames(mat), not_target_col)
#[1] "c" "d"

If we need to select those columns from the matrix

mat[, setdiff(colnames(mat), not_target_col)]

#     c  d
#[1,] 7 10
#[2,] 8 11
#[3,] 9 12

score 0 · Answer 2 · answered Mar 20 '17 at 06:56

0

Another option is %in%

mat[, !colnames(mat) %in% not_target_col]
#     c  d
#[1,] 7 10
#[2,] 8 11
#[3,] 9 12

answered Mar 20 '17 at 06:56

akrun

874,273
37
540
662

How can I select columns/rows with the opposite of the column/row names in R?

2 Answers2