Question

When dimnames is currently NULL, is it possible to re-name a matrix’s dimestions one at a time?

For example, this fails:

mtx <- matrix(1:16,4)
dimnames(mtx)[[2]][1] <- 'col1'

with Error in dimnames(mtx)[[2]][1] <- "col1" : 'dimnames' must be a list


However this works:

mtx <- matrix(1:16,4)
dimnames(mtx)[[1]] <- letters[1:4]
dimnames(mtx)[[2]] <- LETTERS[1:4]
dimnames(mtx)[[2]][1] <- 'col1'
dimnames(mtx)[[2]][2] <- 'col2'

My objective is to seperately replace dimnames(mtx)[[2]][1] and dimnames(mtx)[[2]][2] etc … if this is not possible, i can re-write the loop.

---

Thanks folks, I have ended up with the below — I pass the names in via prepend:

mtxNameSticker <- function(mtx, prepend = NULL, MARGIN=2)
{
    if (MARGIN == 1) max <- nrow(mtx) else
        max  <- ncol(mtx)
    if (is.null(prepend)) if (MARGIN == 2) prepend <- 'C' else
        prepend <- 'R'
    if (length(prepend) == 1) prepend <- paste0(prepend, 1:dim(mtx)[[MARGIN]]) 
    dimnames(mtx)[[MARGIN]] <- seq(from=1, by=1, length.out=dim(mtx)[[MARGIN]]) 
    for (i in 1:max){
        dimnames(mtx)[[MARGIN]][i] <- prepend[i]
    }
    return(mtx)
}

Answer 1

For as long as dimnames is NULL and not an appropriate list, you cannot make assignments to it at particular positions. One easy way to create a dummy but complete list of dimnames is to run:

dimnames(mtx) <- lapply(dim(mtx), seq_len)
mtx
#   1 2  3  4
# 1 1 5  9 13
# 2 2 6 10 14
# 3 3 7 11 15
# 4 4 8 12 16

Then, you can make assignments one at a time like you were wishing:

dimnames(mtx)[[2]][1] <- 'col1'
mtx
#   col1 2  3  4
# 1    1 5  9 13
# 2    2 6 10 14
# 3    3 7 11 15
# 4    4 8 12 16