Loop through array in R

Question

I got the following issue:

I have a large array "x" of the dimension dim(x)= (46, 13, 30). I'm trying to compute a new matrix "M"(30,598) which contains basically the first element of every "slice" in the first column, the second element in the second and so on... Eventually, I want to work with vectors (columns of M) containing all the elements of each grid cell.

With the following function I would be able to calculate my matrix M. However, it's not exactly a user-friendly solution considering the large array:

M <- cbind(x[1,1,],x[1,2,],x[1,3,],x[1,4,],x[2,1,],x[2,2,], x[2,3,],x[2,4,]......)

That's why I want to use a for-loop. Unfortunately I can't figure out how.

This is what I tried so far (unsuccessfully):

M  <- matrix(NA, nrow = 6, ncol = 20)
for (i in 1:120){
    M[i] <- cbind(x[,,i])
}

This loop only creates the first vector (first element of every "slice" [1,1,]), than aborts.

Any ideas on how to get this done? Very much appreciate the help!

It would be helpful if you post some dummy data, e.g. a small array with 3 dimensions, so we can help you better. — user3640617
– user3640617, Commented Aug 11, 2017 at 16:01
Like set.seed(1) ; x <- array(rnorm(2*3*5), dim = c(2, 3, 5)) for instance. And a related desired output — Aurèle
– Aurèle, Commented Aug 11, 2017 at 16:02
In my example, is your desired output desired_output <- cbind(x[1,1,],x[1,2,],x[1,3,],x[2,1,],x[2,2,], x[2,3,])? — Aurèle
– Aurèle, Commented Aug 11, 2017 at 16:06

Aurèle · Accepted Answer · 2017-08-11 16:44:06Z

Sample data:

set.seed(1) ; x <- array(rnorm(2*3*5), dim = c(2, 3, 5))

# , , 1
# 
# [,1]       [,2]       [,3]
# [1,] -0.6264538 -0.8356286  0.3295078
# [2,]  0.1836433  1.5952808 -0.8204684
# 
# , , 2
# 
# [,1]       [,2]      [,3]
# [1,] 0.4874291  0.5757814 1.5117812
# [2,] 0.7383247 -0.3053884 0.3898432
# 
# , , 3
# 
# [,1]        [,2]        [,3]
# [1,] -0.6212406  1.12493092 -0.01619026
# [2,] -2.2146999 -0.04493361  0.94383621
# 
# , , 4
# 
# [,1]      [,2]        [,3]
# [1,] 0.8212212 0.9189774  0.07456498
# [2,] 0.5939013 0.7821363 -1.98935170
# 
# , , 5
# 
# [,1]       [,2]       [,3]
# [1,]  0.61982575 -0.1557955 -0.4781501
# [2,] -0.05612874 -1.4707524  0.4179416

Then permutate dimensions, apply() identity (or, equivalently, as.vector or even c, see ?apply) to the 3rd dimension and transpose:

res <- t(apply(aperm(x, c(3, 2, 1)), 1, identity))

#            [,1]       [,2]        [,3]        [,4]        [,5]       [,6]
# [1,] -0.6264538 -0.8356286  0.32950777  0.18364332  1.59528080 -0.8204684
# [2,]  0.4874291  0.5757814  1.51178117  0.73832471 -0.30538839  0.3898432
# [3,] -0.6212406  1.1249309 -0.01619026 -2.21469989 -0.04493361  0.9438362
# [4,]  0.8212212  0.9189774  0.07456498  0.59390132  0.78213630 -1.9893517
# [5,]  0.6198257 -0.1557955 -0.47815006 -0.05612874 -1.47075238  0.4179416

Equivalently, per @Frank's suggestion: res <- matrix(aperm(x, c(3, 2, 1)), dim(x)[3])

Finally:

desired_output <- cbind(x[1,1,],x[1,2,],x[1,3,],x[2,1,],x[2,2,], x[2,3,])
all.equal(res, desired_output)

# [1] TRUE

Collectives™ on Stack Overflow

Loop through array in R

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related