R/long_to_wide_converter.R

In ipmisc: Miscellaneous Functions for Data Cleaning and Analysis

#' @title Converts dataframe from long/tidy to wide format with `NA`s removed
#' @name long_to_wide_converter
#'
#' @description
#'
#' This conversion is helpful mostly for repeated measures design, where
#' removing `NA`s by participant can be a bit tedious.
#'
#' It does not make sense to spread the dataframe to wide format when the
#' measure is not repeated, so if `paired = TRUE`, `spread` argument will be
#' ignored.
#'
#' @param data A dataframe (or a tibble) from which variables specified are to
#'   be taken. Other data types (e.g., matrix,table, array, etc.) will **not**
#'   be accepted.
#' @param x The grouping (or independent) variable from the dataframe `data`. In
#'   case of a repeated measures or within-subjects design, if `subject.id`
#'   argument is not available or not explicitly specified, the function assumes
#'   that the data has already been sorted by such an id by the user and creates
#'   an internal identifier. So if your data is **not** sorted, the results
#'   *can* be inaccurate when there are more than two levels in `x` and there
#'   are `NA`s present. The data is expected to be sorted by user in
#'   subject-1,subject-2, ..., pattern.
#' @param y The response (or outcome or dependent) variable from the
#'   dataframe `data`.
#' @param subject.id Relevant in case of a repeated measures or within-subjects
#'   design (`paired = TRUE`, i.e.), it specifies the subject or repeated
#'   measures identifier. **Important**: Note that if this argument is `NULL`
#'   (which is the default), the function assumes that the data has already been
#'   sorted by such an id by the user and creates an internal identifier. So if
#'   your data is **not** sorted and you leave this argument unspecified, the
#'   results *can* be inaccurate when there are more than two levels in `x` and
#'   there are `NA`s present.
#' @param paired Logical that decides whether the experimental design is
#'   repeated measures/within-subjects or between-subjects. The default is
#'   `FALSE`.
#' @param spread Logical that decides whether the dataframe needs to be
#'   converted from long/tidy to wide (default: `TRUE`). Relevant only if
#'   `paired = TRUE`.
#' @param ... Currently ignored.
#'
#' @importFrom dplyr row_number select mutate group_by ungroup arrange relocate
#' @importFrom dplyr nest_by filter
#' @importFrom tidyr pivot_longer unnest
#'
#' @return A dataframe with `NA`s removed.
#'
#' @examples
#' # for reproducibility
#' library(ipmisc)
#' set.seed(123)
#'
#' # repeated measures design
#' long_to_wide_converter(
#'   data = bugs_long,
#'   x = condition,
#'   y = desire,
#'   subject.id = subject,
#'   paired = TRUE
#' )
#'
#' # independent measures design (spread argument is ignored)
#' long_to_wide_converter(
#'   data = ggplot2::msleep,
#'   x = vore,
#'   y = brainwt,
#'   paired = FALSE,
#'   spread = FALSE
#' )
#' @export

# function body
long_to_wide_converter <- function(data,
                                   x,
                                   y,
                                   subject.id = NULL,
                                   paired = TRUE,
                                   spread = TRUE,
                                   ...) {
  # make sure both quoted and unquoted arguments are allowed
  c(x, y) %<-% c(rlang::ensym(x), rlang::ensym(y))

  # for non-paired data, even if specified, ignore it
  if (isFALSE(paired)) subject.id <- NULL

  # initial cleanup
  data %<>%
    dplyr::select({{ x }}, {{ y }}, rowid = {{ subject.id }}) %>%
    dplyr::mutate({{ x }} := droplevels(as.factor({{ x }}))) %>%
    dplyr::arrange({{ x }})

  # if `subject.id` wasn't provided, create one for internal usage
  if (!"rowid" %in% names(data)) {
    # the row number needs to be assigned for each participant in paired data
    if (isTRUE(paired)) data %<>% dplyr::group_by({{ x }})

    # unique id for each participant
    data %<>% dplyr::mutate(rowid = dplyr::row_number())
  }

  # NA removal
  data %<>%
    dplyr::ungroup(.) %>%
    dplyr::nest_by(rowid, .key = "df") %>%
    dplyr::filter(sum(is.na(df)) == 0) %>%
    tidyr::unnest(cols = c(df))

  # convert to wide?
  if (spread && paired) data %<>% tidyr::pivot_wider(names_from = {{ x }}, values_from = {{ y }})

  # final clean-up
  as_tibble(dplyr::relocate(data, rowid) %>% dplyr::arrange(rowid))
}