R/similar_permutations.R

In aarong1/RshinyHelpers: Collection of innovative snippets with no equivalent (imo)

#' find the distinct _permutations_ of columns
#' @description find the distinct _permutations_ of columns, particularly in graph applications were
#' describing relationships between unordered nodes, or other applications were the order
#' does not matter
#'
#' @param edges dataframe, tibble, The data structure to work on
#' @param to tidy-eval, character The variable to group by
#' @param from The column of values that will collapse into a list.
#' @return The inputs pasted together as a character string.
#' @details The inputs can be anything that casn be input into
#' the paste function.
#' @examples
#' if(!require(dplyr)){library(dplyr)}
#' edges <- data.frame(
#'   to=sample(c('A','B','C','D'),size = 50,replace=TRUE),
#'   from=sample(c('A','B','C','D','E','F'),size = 50,replace=TRUE))
#'
#' dplyr::count(edges,to,from)
#'
#' ordered = similar_permutations(edges,to,from)
#'
#' dplyr::count(ordered, to ,from)
#' @import dplyr
#' @import tidyr
#' @import tibble
#' @import purrr
#'@export
similar_permutations <- function(edges, to, from) {
  x <- edges %>%
    dplyr::rowwise() %>%
    #-----------
  # important as that the selection is from each ROW and col
  # and we are not grabbing the entire field
  #------------
  dplyr::mutate(to_from = list(c({
    {
      to
    }
  }, {
    {
      from
    }
  })))
  
  
  y <-  purrr::map(x$to_from, sort) %>% #sort alphabetically
    as.data.frame() %>% #prep for transpose
    t() %>% #transpose
    as.data.frame() #coerce back to data.frame- doesn't also preserve df
  
  names(y) <- c('to_ordered', 'from_ordered')
  
  orderedxy <- cbind(x, y) %>%
    tibble::rownames_to_column(var = 'rn') %>%
    dplyr::select(-rn)
  
  #we can then get the ordered fields and count them
  orderedxy <- orderedxy %>%
    dplyr::group_by(to_ordered, from_ordered) %>% #,label
    dplyr::add_count() %>%
    dplyr::filter(n > 1) %>%
    dplyr::ungroup() %>%
    dplyr::select(-c(n, to_ordered, from_ordered, to_from))
  
  return(orderedxy)
}