R/merge_edge_counts.R

In LANL-Bioinformatics/MetaComp: EDGE Taxonomy Assignments Visualization

#' @importFrom dplyr select
NULL

#' Merges two or more EDGE-like taxonomical assignments. The input data frames are assumed to
#' have the following columns: LEVEL, TAXA, and COUNT -- these will be used in the merge
#' procedure, all other columns will be ignored.
#'
#' @param assignments A named list of assignments (the list element's name will be used as a
#'                      resulting data frame column name).
#'
#' @return A merged table, which is a data frame whose rows are taxonomical ids and columns are
#'          the input assignments ids.
#'
#' @export
merge_edge_counts <- function(assignments) {

  # fix CRAN notes
  #
  LEVEL <- TAXA <- COUNT <- NULL # fix the CRAN note

  # extract only rows wich correspond to the desired taxonomy level and name the first column
  #
  res <- dplyr::select(assignments[[1]], LEVEL, TAXA, COUNT)
  names(res) <- c(names(res)[1:2], names(assignments)[1])

  # iterate over the rest of the input list whilst merging the resulting table with
  # the current list's element
  #
  if (length(assignments) > 1) {
    for (i in 2:length(assignments)) {
      res  <- base::merge.data.frame(res, dplyr::select(assignments[[i]], LEVEL, TAXA, COUNT),
                    by = c("LEVEL", "TAXA"), all = T)
      names(res) <- c(names(res)[1:(length(names(res)) - 1)], names(assignments)[i])
    }
  }

  # merge produces NAs when there is no corresponding value in to-be-merged column, fix these with
  # zeros
  #
  res[is.na(res)] <- 0

  # and voila
  #
  res
}