R/epi_clean_class_to_factor.R

In AntonioJBT/episcout: Quickly Clean, Explore and Visualise Large Epidemiological Datasets

#' @title Convert column class to factor if unique values below cutoff
#'
#' @description Convert column class to factor if column has less than
#' x unique results. Useful for handling data frames with many columns with some
#' automation for specifying column classes.
#' Assumes that columns with fewer unique values than the cutoff parameter
#' should be read as factors. All column classes other than dates are converted.
#' Assumes that eg integer columns with few values are in fact coded factors.
#'
#' @param df A data frame object with columns to convert to factor
#'
#' @param cutoff_unique A cutoff for the number of unique values present in the
#' column. Column class will be converted if there are less unique values than
#' the cutoff. Default is 10.
#'
#' @return The data frame passed with classes converted.
#'
#' @note Avoids converting dates to factors if class is already Date.
#' Note that lubridate converted dates may return eg "POSIXct" "POSIXt"
#' which return a logical vector > 1 (for each element)
#' and give a warning message (that can be ignored):
#' "... the condition has length > 1 and only the first element will be used"
#'
#' @author Antonio Berlanga-Taylor <\url{https://github.com/AntonioJBT/episcout}>
#'
#' @seealso \code{\link{epi_clean_count_classes}}, \code{\link{epi_clean_cond_date}},
#' \code{\link{epi_clean_cond_chr_fct}}, \code{\link{epi_clean_cond_numeric}}
#'
#' @examples
#' \dontrun{
#' n <- 20
#' df_factor <- data.frame(
#'   var_id = rep(1:(n / 2), each = 2),
#'   var_to_rep = rep(c("Pre", "Post"), n / 2),
#'   x = rnorm(n),
#'   y = rbinom(n, 1, 0.50),
#'   z = rpois(n, 2)
#' )
#' df_factor$date_col <- seq(as.Date("2018/1/1"),
#'   by = "year", length.out = 5
#' ) # nrow(df_factor))
#' # Check the current classes:
#' lapply(df_factor, class)
#' # Check how many unique values within each column:
#' lapply(df_factor, function(x) length(unique(x)))
#' # Check conditions:
#' i <- "date_col"
#' cutoff_unique <- 10
#' # if num. of unique values is less than cut-off:
#' length(unique(df_factor[[i]])) < cutoff_unique # should be TRUE
#' # and the class is not already a date:
#' class(df_factor[[i]]) != "Date" # should be FALSE
#' df_factor <- epi_clean_class_to_factor(df_factor, cutoff_unique = 10)
#' lapply(df_factor, class)
#' df_factor
#' }
#'
#' @export
#'

epi_clean_class_to_factor <- function(df = NULL,
                                      cutoff_unique = 10) {
  for (i in seq_along(df)) {
    if (
      # if num. of unique values is less than cut-off
      length(unique(df[[i]])) < cutoff_unique &&
        # and the class is not already a date:
        !inherits(df[[i]], "Date")
    ) {
      # Convert to factor:
      df[[i]] <- as.factor(df[[i]])
    }
  }
  return(df)
}