Download grouped remote files using friendlier group names #rstatstips

download-groups.r

library(tidyverse)

# based on {searchable} and https://stackoverflow.com/a/64931927/3246758
# stats::setNames is fine too!
invert <- function(x) {
  set_names(names(x), x)
}

# let's map the coded names to our friendly ones!
fruit <- c(
  "apple" = "ABCJ001",
  "banana" = "ABCJ002",
  "orange" = "AKFJ004")
season <- c(
  "summer" = "DJF",
  "autumn" = "MAM",
  "winter" = "JJA",
  "spring" = "SON")

# now let's work out all the possible combos of fruit and season codes
# stats::expand.grid is fine too!
combos <-
  expand_grid(fruit, season) %>%
  # (we're going to rename these coded columns to avoid getting them mixed
  # up with the unexpanded vectors above!)
  rename(fruit_coded = fruit, season_coded = season) %>%
  # now let's recode using the friendly names of the fruits and seasons...
  mutate(
    fruit_friendly = recode(fruit_coded, !!!invert(fruit)),
    season_friendly = recode(season_coded, !!!invert(season))) %>%
  # and now we can build file paths
  mutate(
    remote_path = paste0("/data/2019/sales/", fruit_coded, "-", season_coded, ".csv"),
    local_path =
      file.path("data", paste0(fruit_friendly, "-", season_friendly, ".csv")))
  
combos
# # A tibble: 12 x 6
#    fruit  season fruit_friendly season_friendly remote_path        local_path   
#    <chr>  <chr>  <chr>          <chr>           <chr>              <chr>        
#  1 ABCJ0… DJF    apple          summer          /data/2019/sales/… data/apple-s…
#  2 ABCJ0… MAM    apple          autumn          /data/2019/sales/… data/apple-a…
#  3 ABCJ0… JJA    apple          winter          /data/2019/sales/… data/apple-w…
#  4 ABCJ0… SON    apple          spring          /data/2019/sales/… data/apple-s…
#  5 ABCJ0… DJF    banana         summer          /data/2019/sales/… data/banana-…
#  6 ABCJ0… MAM    banana         autumn          /data/2019/sales/… data/banana-…
#  7 ABCJ0… JJA    banana         winter          /data/2019/sales/… data/banana-…
#  8 ABCJ0… SON    banana         spring          /data/2019/sales/… data/banana-…
#  9 AKFJ0… DJF    orange         summer          /data/2019/sales/… data/orange-…
# 10 AKFJ0… MAM    orange         autumn          /data/2019/sales/… data/orange-…
# 11 AKFJ0… JJA    orange         winter          /data/2019/sales/… data/orange-…
# 12 AKFJ0… SON    orange         spring          /data/2019/sales/… data/orange-…

# and let's see what the paths look like:
combos %>% select(remote_path, local_path)
# # A tibble: 12 x 2
#    remote_path                      local_path            
#    <chr>                            <chr>                 
#  1 /data/2019/sales/ABCJ001-DJF.csv data/apple-summer.csv 
#  2 /data/2019/sales/ABCJ001-MAM.csv data/apple-autumn.csv 
#  3 /data/2019/sales/ABCJ001-JJA.csv data/apple-winter.csv 
#  4 /data/2019/sales/ABCJ001-SON.csv data/apple-spring.csv 
#  5 /data/2019/sales/ABCJ002-DJF.csv data/banana-summer.csv
#  6 /data/2019/sales/ABCJ002-MAM.csv data/banana-autumn.csv
#  7 /data/2019/sales/ABCJ002-JJA.csv data/banana-winter.csv
#  8 /data/2019/sales/ABCJ002-SON.csv data/banana-spring.csv
#  9 /data/2019/sales/AKFJ004-DJF.csv data/orange-summer.csv
# 10 /data/2019/sales/AKFJ004-MAM.csv data/orange-autumn.csv
# 11 /data/2019/sales/AKFJ004-JJA.csv data/orange-winter.csv
# 12 /data/2019/sales/AKFJ004-SON.csv data/orange-spring.csv

# let's make sure our target folders exist before we start trying to download files!
combos %>%
  pull(local_path) %>%
  dirname() %>%
  unique() %>%
  walk(dir.create, recursive = TRUE)

# now, we're ready to download a bunch of files!
# (cheap plug: you could do this safely with {purrr} or {collateral})
combos %>%
  { walk2(.$remote_path, .$local_path, download.file) }