R/specify_missing.R

In incadata: Recognize and Handle Data in Formats Used by Swedish Cancer Centers

#' Specify missing values for a vector
#' 
#' Change specified values to NA
#' 
#' @param x vector
#' @param ... values that should be changed to \code{NA} if found in \code{x}
#' @param default_missing a vector with additional default values to change 
#' to NA. 
#' These are treated the same as \code{...} but are added by default if not 
#' removed. 
#'   A special value \code{"blank"} can be used to indicate all empty strings 
#'   (all
#'   characters matching \code{[:blank:]}, see \link{regex}).
#'   
#' @return x itself but with specified values set to \code{NA}.
#'   
#' @examples
#' x <- sample(100)
#' x[sample(100, 10)] <- 999
#' specify_missing(x, 999)
#' @export
#' @name specify_missing
specify_missing <- 
  function(x, ..., default_missing = c("", NA, "blanks")){
    UseMethod("specify_missing")
}

#' @export
specify_missing.list <- 
  function(x, ..., default_missing = c("", NA, "blanks")){
    lapply(x, specify_missing, ..., default_missing)
}

#' @export
specify_missing.data.frame <- 
  function(x, ..., default_missing = c("", NA, "blanks")){
    x[] <- specify_missing.list(x, ..., default_missing)
    x
}

#' @export
specify_missing.matrix <- 
  function(x, ..., default_missing = c("", NA, "blanks")){
    apply(x, 2, specify_missing, ..., default_missing)
}


#' @export
specify_missing.factor <-
  function(x, ..., default_missing = c("", NA, "blanks")) {
    as.factor(specify_missing(as.character(x), ..., default_missing))
  } 

#' @export
specify_missing.default <-
  function(x, ..., default_missing = c("", NA, "blanks")){
  
    missing <- c(default_missing, list(...))
    
    # Remove completely blank cells
    # We can not simply use trum, since that would also effect non blank cells
    if ("blanks" %in% missing) {
      x <- ifelse(grepl("^[[:blank:]]*$", x), "", x)
    }
    
    x[x %in% missing] <- NA
    x
}