R/checking.R

In KevinRPan/handy: Commonly Used Functions For Handling Data

#' Functions to help with data checks

#' @importFrom magrittr "%>%"
#' @importFrom magrittr "%<>%"

usethis::use_package("dplyr")
usethis::use_package("purrr")
usethis::use_package("tibble")
usethis::use_package("tidyr")
usethis::use_package("magrittr")
usethis::use_package("stringr")


pct_missings_chr <- function(df) {
  df %>%
    purrr::map_df(~ (sum(is.na(.x)) * 100 /nrow(df)) %>% round(2) %>% as.character)
}

#' @title Check variables in the dataset
#'
#' @description
#' Get a snapshot preview of what variables are available and what they look like.
#'
#' @param df              the input dataframe
#' @param num_unique_vals number of unique values to display
#' @param sort_examples   whether first few examples or random
#'
#' @return Dataframe with Unique Values, Percent Missing and Examples
#' @examples
#' df <- data.frame('a' = letters, 'b' = 1:length(letters), 'c' = rep(NA, length(letters)))
#' check_variables(df)
#' @export
check_variables <- function(df,
                            num_unique_vals = 3,
                            sort_examples = FALSE) {
  ## How many unique values do variables take on?
  var_breakdown <- df %>%
    purrr::map_df(~ .x %>% unique %>% length %>% as.character) %>%
    dplyr::bind_rows(df %>% pct_missings_chr) %>%
    dplyr::bind_rows(df %>%
                       purrr::map_df(~ paste(
                         .x %>%
                           unique %>%
                           stats::na.omit() %>%
                           {if(sort_examples) sort(.) else .} %>%
                           head(num_unique_vals),
                         collapse = ", "))) %>%
    dplyr::bind_cols() %>%
    t %>%
    as.data.frame %>%
    magrittr::set_names(c('Unique Values', 'Percent Missing', 'Example Values')) %>%
    dplyr::mutate_all(as.character) %>%
    dplyr::mutate_at(1:2, as.numeric) %>%
    tibble::rownames_to_column('Variable')

  var_breakdown$Variable <- names(df)

  var_breakdown
}

#' @title Spread based on selected variables
#'
#' @description
#' Select and spread data
#'
#' @param df the input dataframe
#' @param group_cols the columns to group by
#' @param spread_col the column to spread by
#' @param spread_val the column values to spread
#' @param fill_val filling missing with what value
#'
#' @return dataframe that is spread based on values
#' @examples
#' df <- data.frame('id' = c(rep('x',2), rep('y',3),'z'), 'g' = letters[1:6], 'v' = 1:6)
#' spread_column(df, 'id','g','v')
#' @export
spread_column <- function(df, group_cols, spread_col, spread_val, fill_val = 0) {
  df %>%
    dplyr::select_(.dots = c(group_cols, spread_col, spread_val)) %>%
    tidyr::spread_(spread_col, spread_val, fill = fill_val)
}