chipotle_cleaning

#week 21
library(tidyverse)
url <- "https://raw.githubusercontent.com/TheUpshot/chipotle/master/orders.tsv"

df <- read_tsv(url)

tidy_df <- df %>% 
  separate(choice_description, c("Filling 1", "Filling 2", "Filling 3", "Filling 4", "Filling 5", "Filling 6", 
                                 "Filling 7", "Filling 8", "Filling 9", "Filling 10"), 
           extra = "merge",
           fill = "right",
           sep = ",") %>%
  mutate(`Meal item` = case_when(str_detect(tolower(item_name), "burrito|bowl|salad|taco") ~ item_name,
                               TRUE ~ NA_character_)) %>% 
  select(order_id:item_name, item_price, `Meal item`, everything()) %>% 
  gather(key = "ingredient_number", value = "ingredient", `Meal item`:`Filling 10`) %>%
  mutate(
    #ingredient_number = case_when(str_detect(tolower(item_name), "burrito|bowl|salad|taco") ~ ingredient_number,
                                      # TRUE ~ str_extract(ingredient_number, "[:digit:]+")),
         ingredient = str_remove_all(ingredient, "\\[|\\]"),
         chips = case_when(str_detect(item_name, "Chips and") ~ item_name,
                           TRUE ~ "NOPE"),
         ingredient = case_when(ingredient == "NULL" ~ chips,
                                TRUE ~ ingredient),
         ingredient = case_when(ingredient == "NOPE" ~ item_name,
                                TRUE ~ ingredient)) %>% 
  select(-chips) %>% 
  na.omit()