Using data.table in R, can I select a vector of variable columns from a table?

Question

Suppose I have a table like:

> A = data.table(c(1,2,3),c(4,5,6),c(7,8,9))
> A
    V1 V2 V3
1:  1  4  7
2:  2  5  8
3:  3  6  9

I want to select the first column of the first row, second of the second row and third of the third row, getting the result:

> B
[1] 1 5 9

I've tried something like:

B = A[,c(1,2,3)]

But that does not work. Is it possible to do this? (Obviously in my actual use case the column indexes for each row are going to be variables, but I can guarantee a vector of length equal to the number of rows in the table)

works for this specific example, but what if the columns I want are c(1,1,2)? — Sinnombre
– Sinnombre, Commented Sep 19, 2019 at 20:40

akrun · Accepted Answer · 2019-09-19 19:56:13Z

2

A vectorized way of subsetting the elements is by providing the row index and column index as a matrix

as.data.frame(A)[cbind(seq_len(nrow(A)), 1:3)]
#[1] 1 5 9

Or convert the .SD to matrix or data.frame and use the row/column index

A[, as.matrix(.SD)[cbind(1:3, 1:3)]]

Or in data.table, pass the i, j index in a loop and extract the elements

A[, unlist(Map(function(i, j) .SD[i, j, with = FALSE], 1:3, 1:3), 
          use.names = FALSE)]

edited Sep 19, 2019 at 19:56

answered Sep 19, 2019 at 19:13

akrun

891k38 gold badges590 silver badges700 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Sinnombre Over a year ago

last one is exactly what I need, thanks! To generalize for anyone looking for this later, the first 1:3 in the @akrun solution is a vector of row ids and the second is a vector of column ids. Both can be replaced by lists, and can feature repetitions or not include every vector, but they must be the same length. For example, A[, unlist(Map(function(i, j) .SD[i, j, with = FALSE], c(2,2,1,3), c(2,3,2,1)), use.names = FALSE)] returns: [1] 5 8 4 3

Collectives™ on Stack Overflow

Using data.table in R, can I select a vector of variable columns from a table?

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related