R/dwd_clean.R

In malinfischer/phenoTS: What the Package Does (One Line, Title Case)

#' @title Clean observation data
#'
#' @description This function allows you to clean observation data which is already joined with station and phase information by filtering relevant information. Duplicates are removed and irrelevant columns as well as incomplete or low-quality (based on quality flags) observations are deleted.
#'
#' @param dwd_data tibble containing DWD observation data to which station and phase information has been added already = return of \code{\link{dwd_add_phase_info}} and/or \code{\link{dwd_add_station_info}}.
#'
#' @return A tidyverse tibble containing the cleaned observation data.
#'
#' @import tidyverse
#'
#' @export
#'
#' @examples
#' # read file into R
#' my_file <- dwd_read(dir_tmp) # result: tidyverse tibble
#'
#' # add phase information from matching phase meta file
#' my_file <- dwd_add_phase_info(my_file)
#'
#' # add station information from matching station meta file
#' my_file <- dwd_add_station_info(my_file)
#'
#' # clean observation data
#' my_file <- dwd_clean(my_file)
#'

dwd_clean <- function(dwd_data){

  # remove (full) duplicates as time spans of JM and SM might overlap (usually 1 year)
  cleaned_data <- dplyr::distinct(dwd_data)

  # remove duplicates with different quality information although same observation - only look at "characteristic" columns
  cleaned_data <- dplyr::distinct(cleaned_data,stat_id,ref_year,phase_id,.keep_all = T)

  # remove column "object" which only contains (always similar) name of observed crop (from phase info file)
  cleaned_data <- dplyr::select(cleaned_data,-"obj")

  # remove column "dir" which only contains (always similar) file directory
  cleaned_data <- dplyr::select(cleaned_data,-"dir")

  # remove observations without any station information
  cleaned_data <- dplyr::filter(cleaned_data,!is.na(stat_name))

  # remove low-quality observations based on attached quality byte ("Qualitaetsbyte" 5 (doubtful) or 8 (wrong))
  cleaned_data <- dplyr::filter(cleaned_data,entry_qual!=5)
  cleaned_data <- dplyr::filter(cleaned_data,entry_qual!=8)

  return(cleaned_data)

}