R/to.dummy.R

In varhandle: Functions for Robust Variable Handling

## Function Description:
##     This function turns a categorical character column into multiple binary
##     columns. eg: change the gender (male/female) into male column and female
##     column.

to.dummy <- function(v = NULL, prefix = NULL){
    #----[ checking the input ]----#
    {
        # check if user has not provided prefix argument
        if (is.null(prefix)) {
            stop("The input \"prefix\" is missing. This will be added to the begining of each column name to avoid conflicts with other column names.")
        }else if (length(prefix) != 1 | nchar(prefix) == 0 | !inherits(x = prefix, what = "character")) {
            stop("The input \"prefix\" should be a character vector with length of 1 and character number more than 0.")
        }


        # check if user has not provided v argument
        if (is.null(v)) {
            stop("The input \"v\" is missing. It should be a vector with categories that you are willing to create dummy variables from. It can be a factor, character or numeric vector.")
        }
    }


    #----[ pre-processing ]----#
    {
        ## convert to character vector if the input is factor
        if (inherits(v, "factor")) {
            # get the levels of v
            v_levels <- levels(v)
            # convert the factor to character vector
            v <- as.character(v)
        }else{
            # find tha NAs and turn them into character to have a separate column for NAs
            v[which(is.na(v))] <- "NA"
            # get the categories
            v_levels <- names(table(v, useNA = "ifany"))
        }
    }


    #----[ processing ]----#
    {
        # go through categories one by one
        for (i in 1:length(v_levels)) { 
            # create a logical vector which has 1 for places that has the category
            assign(x = paste("v", i, sep = ""), value = as.numeric(v == v_levels[i]))
        }


        # create a cbind command and run it. It attaches the variables generated in the for loop above.
        df <- eval(parse(text = paste("cbind(",
                                      paste('v', 1:i,
                                            sep = '', collapse = ", "),
                                      ")", collapse = "", sep = "")))
        
        # strip the white space from begining and end of the name and the middle white space with "_"
        factor_levels <- gsub("\\s+", "_", gsub("^\\s+|\\s+$", "", v_levels))
        # if one of the levels are "", we should create a name for it, so we use "BLANK"
        factor_levels[which(factor_levels == "")] <- "BLANK"
        # set the colnames
        colnames(df) <- paste(prefix, factor_levels, sep = ".")
        # return the final data.frame
        return(df)
    }
}