R/humanizeVariable.R

In Eucrow/sapmuebase: Base utils to use in SAP MUE

# ---- function to humanize one variable
#' Humanize variable
#
#' Add new column with the description of a variable coded... I mean, of a
#' encoded column create a new one with the appropriate description.
#' It's available for variables with a data source: COD_PUERTO, COD_BARCO,
#' COD_ARTE, COD_ORIGEN, COD_TIPO_MUE, COD_ESP_MUE, COD_ESP_CAT.
#' Before the humanization, checks the correct format of the variable.
#'
#' @param df df which contains the variable to humanize
#' @param variable one of this values: COD_PUERTO, COD_BARCO,
#' COD_ARTE, COD_ORIGEN, COD_TIPO_MUE, COD_ESP_MUE, COD_ESP_CAT
#' @return Return the df data frame with the new variable added.
#' @export

# TO DO: Add CODSGPM. In maestro_flota_sireno dataset, one CODSGPM can be in various
# rows.
humanizeVariable <- function(df, variable){

  #check if variable exists in the data frame
  if(variable %in% colnames(df)){

    variable_data <- variables_to_humanize[variables_to_humanize[["ORIGINAL_VAR"]] %in% variable,]

    # check that there are just one record for the variable to humanize in variables_to_humanize dataset
    if (nrow(variable_data) > 1) {
      stop (paste("There is an error with the", variable, "variable.
                  Check if the variable exists in variables_to_humanize dataset or if
                  there are various records for", variable, "."))
    } else if (nrow(variable_data) == 0) {

      stop (paste(variable, " is not a variable available to humanize."))

    } else {

      # check the correct format of the variable
      result <- tryCatch({

        checkFormatVariable(df, variable)

      }, error = function(err){

        stop(err)

      }
      )

      # original_var is the variable argument in the function, but is extracted here
      # from variable_data for a better understanding
      original_var <- as.character(variable_data[["ORIGINAL_VAR"]])
      master_var <- as.character(variable_data[["MASTER_VAR"]])
      humanized_master_var <- as.character(variable_data[["HUMANIZED_MASTER_VAR"]])
      humanized_var <- as.character(variable_data[["HUMANIZED_VAR"]])
      master <- as.character(variable_data[["MASTER"]])

      # this is the dataset to join with the dataframe to humanize
      dataframe_to_join_with <- get(master)[, c(master_var, humanized_master_var)]

      # vector of with original column names (usefull to reorder columns after merge)
      column_names_df <- colnames(df)

      if (humanized_var %in% column_names_df){

        warning(paste(humanized_var, "already exists"))
        return(df)

      } else {

        df <- merge(df, dataframe_to_join_with, by.x = original_var, by.y = master_var, all.x = TRUE )
        # this does not work:
        # df <- left_join(df, dataframe_to_join_with, by = c(original_var, master_var))
        # so we have to change the column order:
        # put new column at the end of the dataframe
        df <- df %>%
          select(one_of(column_names_df), everything())
        # and reorder (remember, the new humanized column is the last one, and the
        # variable to humanize is the first one):

        # index of the variable to humanize (usefull to reorder columns after merge)
        index_variable_to_humanize <- which(column_names_df == original_var)

        # reorder columns
        df <- df %>%
          select(1:index_variable_to_humanize, ncol(df), everything())

        # change the name of the new column, because in some cases, the final column
        # name is different of the humanized variable in master
        colnames(df)[which(names(df)==humanized_master_var)] <- humanized_var

        return (df)
      }

    }
  } else {

    stop(paste("Variable", variable, "does not exists in dataframe"))

  }
}