CODEWITHSUNDEEP
arraysrmatrixapply
pdevar
pdevar

Reputation: 333

R Apply Array Output Dimension

Related to this question: Multiply multidimensional array with same-sized matrix

If I create the following array and matrix:

a <- array(1, dim=c(2,2,3))
b <- matrix(c(1,1,1,1), nrow=2)

and would like to carray out an elementwise multiplication of b with each slice of a moving along the 3rd dimension, I would use apply. However, I get the following strange result:

> dim(apply(a, 3, `*`, b))
[1] 4 3
> newa <- array(0, dim=c(2,2,3))
> newa[] <- apply(a, 3, `*`, b)
> dim(newa)
[1] 2 2 3

Why do these two things give different answers?

Upvotes: 2

Views: 639

Answers (1)

aichao
aichao

Reputation: 7455

The resulting dimension of apply(a, 3,*, b) is 4 3 because according to the documentation, see ?apply:

If each call to FUN returns a vector of length n, then apply returns an array of dimension c(n, dim(X)[MARGIN]) if n > 1.

In this case, each call to FUN returns a vector of length 4 and the dimension with MARGIN=3 is 3. Note that the results of each call of FUN in apply is coerced to a vector not a matrix.

Now, with newa[] <- array(..., you are replacing the values in the newa object with the values from the array computation. The attributes of newa is not changed, so its dimensions is still 2 2 3.

Hope this helps.

Upvotes: 2

Related Questions