R/getMissingRLE.R

In bocinsky/guedesbocinsky2018: Research Compendium for d'Alpoim Guedes and Bocinsky (2018)

#' Get the run length encoding of joint NAs in the climate records.
#'
#' Useful for deciding the cutoff value for the number of contiguous days.
#'
#' @param vect A numebric vector
#'
#' @return the RLE of the vector
#' @export
getMissingRLE <- function(vect) {
  ## Read in both records
  vect.tmin <- vect[["TMIN"]]
  vect.tmax <- vect[["TMAX"]]

  ## Make sure they are both the same dimensions
  if (!R.oo::equals(dim(vect.tmin), dim(vect.tmax))) {
    return(NULL)
  }

  ## Get just the climate info
  vect.tmin.records <- as.matrix(vect.tmin[, 3:33])
  vect.tmax.records <- as.matrix(vect.tmax[, 3:33])

  ## Remove all records where we don't have both data
  vect.tmin.records[is.na(vect.tmax.records)] <- NA
  vect.tmax.records[is.na(vect.tmin.records)] <- NA

  ## Get the number of days per month in the records
  n.days <- Hmisc::monthDays(as.Date(paste(vect.tmin$YEAR, vect.tmin$MONTH, "01", sep = "-")))

  ## Unnwrap each row, accounting for number of days in the month, and count
  ## consecutive missing values (these are the same)
  vect.tmin.records.rle <- rle(is.na(unwrapRows(vect.tmin.records, n.days)))
  #   vect.tmax.records.rle <- rle(is.na(unwrap.rows(vect.tmax.records,n.days)))

  out <- vect.tmin.records.rle$lengths[vect.tmin.records.rle$values]
  return(out)
}