anonymise.R

# returns A-Z or AA-ZZ for up to 26*26=676 unique categories
anonymise <- function(mycats) {
  cnt <- length(unique(mycats))
  mycats <- factor(mycats, labels=1:cnt)
  if (cnt <= 26)
    return(LETTERS[mycats])
  newcats <- character(0)
  for (i in 1:ceiling(cnt/26))
    newcats <- c(newcats, paste0(LETTERS[i],LETTERS[1:26]))
  return(newcats[mycats])
}

# test use case
x1 <- sample(c(letters[1:26],LETTERS[1:26]), 100, replace=T)
anonymise(x1)
anonymise(c(letters[1:26], LETTERS[1:26], "A","a"))