R/column_based_recode.R

In mde: Missing Data Explorer

#' Recode NA as based on Other Columns
#' @param  values_from Character. Name of column to get the original values from
#' @param criterion Currently supports one of all_na or any_na to index rows that are either all NA or contain any NA.
#' @param values_to  Character New column name for the newly recoded values. Defaults to the same name if none is supplied.
#' @inheritParams recode_as_na
#' @title Conditionally Recode NA values based on other Columns
#' @return A `data.frame` object with target `NA` values replaced.
#' @examples
#' df <- structure(list(id = 40:43, v1 = c(NA, 1L, 1L, 1L), v2 = c(NA, 1L, 1L, 1L),
#' v3 = c(NA, 2L, NA, 1L),
#' test = c(1L, 2L, 1L, 3L)), class = "data.frame", row.names = c(NA, -4L))
#' # recode test as 0 if all NA, return test otherwise
#' column_based_recode(df,values_from = "test", pattern_type = "starts_with", pattern="v")
#' @export

column_based_recode <- function(df, criterion="all_na", values_from=NULL,
                                values_to=NULL, value=0, pattern_type="contains",pattern="Solar",
                                case_sensitive = FALSE){
  UseMethod("column_based_recode")

}

#' @export

column_based_recode.data.frame <- function(df, criterion="all_na", values_from=NULL,
                                values_to=NULL, value=0, pattern_type="contains",pattern="Solar",
                                case_sensitive = FALSE){

if(is.null(values_to)) values_to = values_from

stopifnot("criterion should be one of all_na or any_na for now" = criterion %in% c("all_na", "any_na"))

use_criterion = all_na.default

if(criterion == "any_na") use_criterion = anyNA

use_these_columns <- recode_selectors(df, pattern_type = pattern_type, pattern = pattern, case_sensitive = case_sensitive)

df %>%
    mutate(!!values_to := ifelse(apply(df[use_these_columns],1,use_criterion), value, !!dplyr::sym(values_from)))


}