R/ind2arr.ind.R

In arnejohannesholmin/TSD: Read and write in the Time Step Data (TSD) format

#*********************************************
#*********************************************
#' Transforms vector indexes to array indexes. The function is cloned from the function which().
#'
#' @param ind	The index(es) to be transformed to array indexes.
#' @param shape	The dimension of the array.
#' @param perm	The permutation i. e. the order of the dimensions along which the indexes given in 'ind' are organized. If 'shape' is 2:4 and 'perm' is c(3,1,2), the indexes in 'ind' are first organized along the third dimension, then along the first, and finally along the second dimension of a matrix of dimensions 2:4.
#' @param ...	Used for robustness.
#'
#' @return
#'
#' @examples
#' \dontrun{}
#'
#' @export
#' @rdname ind2arr.ind
#'
ind2arr.ind <- function(ind, shape=NULL, perm=NULL, ...){
	
	############### LOG: ###############
	# Start: 2009-05-18 - Function cloned from the function which().
	# Last: 2009-05-21 - Added support for 'perm'.
		
	# If the input array indices are empty, return:
	if(length(ind)==0){
		return()
	}
	if(!is.null(shape)){
		# Dimnesions of the imaginary matrix:
		m <- length(ind)
		rank <- length(shape)
		rankseq <- 1:rank
		# See description of 'perm':
		if(!all(perm%in%rankseq)){
			stop(paste("The elements of 'perm' need to be in the range [1:",rank,"]"),"")
		}
		if(is.null(perm)){
			perm <- rankseq
		}
		# Remove 1 here and add 1 to the end to avoid zeros in the output:
		ind1 <- ind - 1
		ind <- 1 + ind1%%shape[perm[1]]
		restperm=perm[-1]
		lastdim=shape[perm[1]]
		ind <- matrix(ind, nrow = m, ncol = rank)
		# Adding colnames to 'ind':
		if (rank == 2){
			colnames(ind)=c("row", "col")
		}
		else{
			colnames(ind)=paste("shape", 1:rank, sep = "")
		}
		if (rank >= 2) {
			denom <- 1
			for (i in 2:rank) {
				denom <- denom * shape[perm[i-1]]
				nextd1 <- ind1%/%denom
				# Array indexes are put along the proper dimension as given by 'perm':
				ind[, perm[i]] <- 1 + nextd1%%shape[perm[i]]
			}
		}
	}
	ind
}