In R, how to index only the last dimension of an array when you don't know how many dimensions are there?

Question

I have an array x that I don't know its dimension in advance. How do I index only the last dimension? For example:

trimGibbsChain <- function (x, ntrim) {
    toTrim <- -(1:ntrim)
    return(x[,,,toTrim]) # How many commas?
}

The problem is, when you don't know the dimension of the array, you don't know how many commas to put there. What is the fastest way to do this?

You might find something [here](http://stackoverflow.com/questions/10865337/r-how-to-get-a-value-of-a-multi-dimensional-array-by-a-vector-of-indices) and [here](http://stackoverflow.com/questions/17750893/how-to-pass-nothing-as-an-argument-to-for-subsetting) — alexis_laz, May 21 '17 at 17:45

mt1022 · Accepted Answer · 2017-05-21T18:44:24.333

Solution based on links in alexis_laz's comment

This can be done by subsetting with [ using list as indices:

lastInd <- function(x, n){
    nd <- length(dim(x))
    # uncomment the following line if x could be a plain vector without dim
    # if(nd == 0) nd = 1
    inds <- rep(alist(,)[1], nd)
    inds[nd] <- n
    do.call(`[`, c(list(x), inds))
}

Original answer

Couldn't find one existing function, but the following should work: x is an array with at least 2-dim; n is the index along the last dim;

lastInd <- function(x, n){
    d <- dim(x)
    d.new <- d[-length(d)]
    block.size <- prod(d.new)
    res <- x[(block.size * (n - 1) + 1):(block.size * n)]
    array(res, dim = d.new)
}

To handle the case when x is a plain vector:

lastInd <- function(x, n){
    d <- dim(x)
    if(is.null(d)){
        x[n]
    }else{
        d.new <- d[-length(d)]
        block.size <- prod(d.new)
        res <- x[(block.size * (n - 1) + 1):(block.size * n)]
        array(res, dim = d.new)
    }
}

An example:

x <- array(1:12, dim = c(2, 2, 3))
lastInd(x, 2)

#      [,1] [,2]
# [1,]    5    7
# [2,]    6    8

lastInd(2:4, 3)
# [1] 4

In R, how to index only the last dimension of an array when you don't know how many dimensions are there?

1 Answers1

Solution based on links in alexis_laz's comment

Original answer