Question

Say, I have a matrix m with 2 rows and 3 columns, together with a function which takes in an argument of numeric type and return a vector. For example,

f <- function(x){
    rep(x,4)
}

If I use

 t <- apply(m, MARGIN = c(1,2), FUN = f)

I will get a tensor t with dim(t) == c(4,2,3). But what I want is a tensor t with dim(t) == c(2,3,4). How can I do this conveniently? Thank you.

Answer 1

Use aperm, a generalized transposition:

tt <- <- apply(m, MARGIN = c(1,2), FUN = f)
aperm(tt, c(2,3,1))

See here: How to change order of array dimensions