R/03_merge_all.R

In aidR: Reading and Standardizing Automated Insulin Delivery Data

##################################
#                                #
#   Merge data across all ids    #
#                                #
##################################

#------------------------
# Public functions
#------------------------

#' Merge data available for all individuals, separately per data type
#'
#' @param data A named list with the data of each individual (the output of
#'   \code{\link{parse_data}}), named by participant identifier.
#' @param types A character vector with the data type names to be merged.
#'   Defaults to the five standardized types and the two daily insulin totals.
#'
#' @return A named list with one data frame per data type, holding the data of
#'   all individuals. Individuals without data of a given type contribute no rows.
#' @examples
#' all_data <- list(
#'   "1" = parse_data("1", system.file("extdata", "glooko_example.zip", package = "aidR")),
#'   "2" = parse_data("2", system.file("extdata", "dexcom_clarity_example.csv", package = "aidR"))
#' )
#' combined <- merge_all(all_data)
#' names(combined)
#' head(combined$cgm)
#' table(combined$cgm$id)
#'
#' # Merge CGM data only
#' merge_all(all_data, types = "cgm")
#' @export
merge_all <- function(data, types = c(
                        "cgm", "basal", "bolus", "carbs", "SMBG",
                        "total_basal", "total_bolus"
                      )) {
  # Merge: CGM, basal, bolus, carbs and SMBG (standardized formats)

  result <- list()
  for (type in types) {
    type_list <- list()

    for (id in names(data)) {
      df <- data[[id]][[type]]
      if (is.null(df)) next # type doesn't exist for this id
      type_list[[id]] <- df %>%
        mutate(
          across(where(is.character), \(x) type.convert(x, as.is = TRUE)), # to avoid type-mismatches
          id = as.character(id) # to ensure consistent ID type across subjects
        )
    }

    result[[type]] <- bind_rows(type_list)
  }

  return(result)
}