Travis-CI Build Status CRAN_Status_Badge

repurrrsive

The repurrrsive package provides recursive lists that are handy when teaching or exampling functions such as purrr::map() and the unnest_*() functions in the tidyr package. Datasets are stored as R list, JSON, and XML to provide the full non-rectangular data experience. Enjoy!

Package also includes the main data frame from the gapminder package in 3 different forms: simple data frame (no list-columns), data frame nested by country, and split into a named list of data frames.

Resources that use repurrrsive, where you can much more list wrangling:

Installation

You can install repurrrsive from CRAN like so:

install.packages("repurrrsive")

or from GitHub with:

# install.packages("devtools")
devtools::install_github("jennybc/repurrrsive")

Recursive list examples

Game of Thrones POV characters

got_chars is a list with information on the 30 point-of-view characters from the first five books in the Song of Ice and Fire series by George R. R. Martin. Retrieved from An API Of Ice And Fire.

library(repurrrsive)
library(purrr)
(nms <- map_chr(got_chars, "name"))
#>  [1] "Theon Greyjoy"      "Tyrion Lannister"   "Victarion Greyjoy" 
#>  [4] "Will"               "Areo Hotah"         "Chett"             
#>  [7] "Cressen"            "Arianne Martell"    "Daenerys Targaryen"
#> [10] "Davos Seaworth"     "Arya Stark"         "Arys Oakheart"     
#> [13] "Asha Greyjoy"       "Barristan Selmy"    "Varamyr"           
#> [16] "Brandon Stark"      "Brienne of Tarth"   "Catelyn Stark"     
#> [19] "Cersei Lannister"   "Eddard Stark"       "Jaime Lannister"   
#> [22] "Jon Connington"     "Jon Snow"           "Aeron Greyjoy"     
#> [25] "Kevan Lannister"    "Melisandre"         "Merrett Frey"      
#> [28] "Quentyn Martell"    "Samwell Tarly"      "Sansa Stark"
map_dfr(got_chars, `[`, c("name", "gender", "culture", "born"))
#> # A tibble: 30 x 4
#>    name               gender culture  born                                 
#>    <chr>              <chr>  <chr>    <chr>                                
#>  1 Theon Greyjoy      Male   Ironborn In 278 AC or 279 AC, at Pyke         
#>  2 Tyrion Lannister   Male   ""       In 273 AC, at Casterly Rock          
#>  3 Victarion Greyjoy  Male   Ironborn In 268 AC or before, at Pyke         
#>  4 Will               Male   ""       ""                                   
#>  5 Areo Hotah         Male   Norvoshi In 257 AC or before, at Norvos       
#>  6 Chett              Male   ""       At Hag's Mire                        
#>  7 Cressen            Male   ""       In 219 AC or 220 AC                  
#>  8 Arianne Martell    Female Dornish  In 276 AC, at Sunspear               
#>  9 Daenerys Targaryen Female Valyrian In 284 AC, at Dragonstone            
#> 10 Davos Seaworth     Male   Westeros In 260 AC or before, at King's Landi…
#> # … with 20 more rows

The same got_chars data is also present as JSON and XML files. Accessor functions provide the local file path.

got_chars_json()
#> [1] "/Users/jenny/resources/R/library_3.6/repurrrsive/extdata/got_chars.json"
got_chars_xml()
#> [1] "/Users/jenny/resources/R/library_3.6/repurrrsive/extdata/got_chars.xml"

Practice bringing data from JSON into an R list.

library(jsonlite)
json <- fromJSON(got_chars_json(), simplifyDataFrame = FALSE)
json[[1]][c("name", "titles", "playedBy")]
#> $name
#> [1] "Theon Greyjoy"
#> 
#> $titles
#> [1] "Prince of Winterfell"                                
#> [2] "Captain of Sea Bitch"                                
#> [3] "Lord of the Iron Islands (by law of the green lands)"
#> 
#> $playedBy
#> [1] "Alfie Allen"
identical(got_chars, json)
#> [1] TRUE

Practice bringing data into R from XML. You can get it into an R list with xml2::as_list(), but to get a list as nice as those above? That requires a bit more work. Such is XML life.

library(xml2)
xml <- read_xml(got_chars_xml())
xml_child(xml)
#> {xml_node}
#> <elem>
#>  [1] <url>https://www.anapioficeandfire.com/api/characters/1022</url>
#>  [2] <id>1022</id>
#>  [3] <name>Theon Greyjoy</name>
#>  [4] <gender>Male</gender>
#>  [5] <culture>Ironborn</culture>
#>  [6] <born>In 278 AC or 279 AC, at Pyke</born>
#>  [7] <died/>
#>  [8] <alive>TRUE</alive>
#>  [9] <titles>\n  <elem>Prince of Winterfell</elem>\n  <elem>Captain of S ...
#> [10] <aliases>\n  <elem>Prince of Fools</elem>\n  <elem>Theon Turncloak< ...
#> [11] <father/>
#> [12] <mother/>
#> [13] <spouse/>
#> [14] <allegiances>House Greyjoy of Pyke</allegiances>
#> [15] <books>\n  <elem>A Game of Thrones</elem>\n  <elem>A Storm of Sword ...
#> [16] <povBooks>\n  <elem>A Clash of Kings</elem>\n  <elem>A Dance with D ...
#> [17] <tvSeries>\n  <elem>Season 1</elem>\n  <elem>Season 2</elem>\n  <el ...
#> [18] <playedBy>Alfie Allen</playedBy>

Star Wars Universe entities

sw_people, sw_films, sw_species, sw_planets, sw_starships and sw_vehicles are interrelated lists about entities in the Star Wars Universe retrieved from the Star Wars API using the package rwars.

library(repurrrsive)
library(purrr)
map_chr(sw_films, "title") 
#> [1] "A New Hope"              "Attack of the Clones"   
#> [3] "The Phantom Menace"      "Revenge of the Sith"    
#> [5] "Return of the Jedi"      "The Empire Strikes Back"
#> [7] "The Force Awakens"

Elements that contain URLs provide a way to link the lists together. For example, the films element of each person contains URLs for the films they have appeared in. For example, Luke Skywalker has been in five films.

luke <- sw_people[[1]]
names(luke)
#>  [1] "name"       "height"     "mass"       "hair_color" "skin_color"
#>  [6] "eye_color"  "birth_year" "gender"     "homeworld"  "films"     
#> [11] "species"    "vehicles"   "starships"  "created"    "edited"    
#> [16] "url"
luke[["films"]]
#> [1] "http://swapi.co/api/films/6/" "http://swapi.co/api/films/3/"
#> [3] "http://swapi.co/api/films/2/" "http://swapi.co/api/films/1/"
#> [5] "http://swapi.co/api/films/7/"

These URLs can be looked up in the the sw_films list to find the titles of the films.

# Create a mapping between titles and urls
film_lookup <- map_chr(sw_films, "title") %>% 
  set_names(map_chr(sw_films, "url"))

# The films Luke is in
film_lookup[luke[["films"]]] %>% unname()
#> [1] "Revenge of the Sith"     "Return of the Jedi"     
#> [3] "The Empire Strikes Back" "A New Hope"             
#> [5] "The Force Awakens"

GitHub user and repo data

gh_users and gh_repos are lists with information for 6 GitHub users and up to 30 of each user’s repositories.

GitHub users.

library(repurrrsive)
library(purrr)
map_chr(gh_users, "login")
#> [1] "gaborcsardi" "jennybc"     "jtleek"      "juliasilge"  "leeper"     
#> [6] "masalmon"
map_chr(gh_users, 18)
#> [1] "Gábor Csárdi"           "Jennifer (Jenny) Bryan"
#> [3] "Jeff L."                "Julia Silge"           
#> [5] "Thomas J. Leeper"       "Maëlle Salmon"
map_dfr(gh_users, `[`, c("login", "name", "id", "location"))
#> # A tibble: 6 x 4
#>   login       name                         id location              
#>   <chr>       <chr>                     <int> <chr>                 
#> 1 gaborcsardi Gábor Csárdi             660288 Chippenham, UK        
#> 2 jennybc     Jennifer (Jenny) Bryan   599454 Vancouver, BC, Canada 
#> 3 jtleek      Jeff L.                 1571674 Baltimore,MD          
#> 4 juliasilge  Julia Silge            12505835 Salt Lake City, UT    
#> 5 leeper      Thomas J. Leeper        3505428 London, United Kingdom
#> 6 masalmon    Maëlle Salmon           8360597 Barcelona, Spain

First ~30 repos of these users. Peek at some info from first repo for the first user. Get full name of each user’s 11th repo.

str(gh_repos[[1]][[1]][c("full_name", "html_url", "description")])
#> List of 3
#>  $ full_name  : chr "gaborcsardi/after"
#>  $ html_url   : chr "https://github.com/gaborcsardi/after"
#>  $ description: chr "Run Code in the Background"
map_chr(gh_repos, list(11, "full_name"))
#> [1] "gaborcsardi/debugme"                     
#> [2] "jennybc/access-r-source"                 
#> [3] "jtleek/datawomenontwitter"               
#> [4] "juliasilge/juliasilge.github.io"         
#> [5] "leeper/congressional-district-boundaries"
#> [6] "masalmon/geoparsing_tweets"

Want to parse it yourself? Paths to local JSON and XML files.

c(gh_users_json(), gh_repos_json(), gh_users_xml(), gh_repos_xml())
#> [1] "/Users/jenny/resources/R/library_3.6/repurrrsive/extdata/gh_users.json"
#> [2] "/Users/jenny/resources/R/library_3.6/repurrrsive/extdata/gh_repos.json"
#> [3] "/Users/jenny/resources/R/library_3.6/repurrrsive/extdata/gh_users.xml" 
#> [4] "/Users/jenny/resources/R/library_3.6/repurrrsive/extdata/gh_repos.xml"

Redo this: Get full name of each user’s 11th repo. But using only the XML.

library(xml2)
repo_xml <- read_xml(gh_repos_xml())
repo_names <- map_chr(xml_find_all(repo_xml, "//full_name"), xml_text)
elevenses <- 
  11 + cumsum(c(0, head(table(gsub("(.*)/.*", "\\1", repo_names)), -1)))
repo_names[elevenses]
#> [1] "gaborcsardi/debugme"                     
#> [2] "jennybc/access-r-source"                 
#> [3] "jtleek/datawomenontwitter"               
#> [4] "juliasilge/juliasilge.github.io"         
#> [5] "leeper/congressional-district-boundaries"
#> [6] "masalmon/geoparsing_tweets"

Sharla Gelfand’s music collection

discog holds a list of 155 items, representing a music collection stored in the Discogs database and retrieved via their API. It’s useful for demonstrating capabilities of purrr and tidyr.

library(repurrrsive)
library(purrr)
library(tidyr) # version >= 0.8.3.9000
library(tibble)
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union

Here we get just the album names:

discog %>% 
  map_chr(c("basic_information", "title")) %>% 
  head()
#> [1] "Demo"                              "Observant Com El Mon Es Destrueix"
#> [3] "I"                                 "Oído Absoluto"                    
#> [5] "A Cat's Cause, No Dogs Problem"    "Tashme"

Put the discog list into a list-column and use tidyr::hoist() to dig some info out of it.

tibble(disc = discog) %>% 
  hoist(disc, info = "basic_information") %>% 
  hoist(info,
       title = "title",
       year = "year",
       label = list("labels", 1, "name"),
       artist = list("artists", 1, "name")
  ) %>% 
  select(-disc, -info)
#> # A tibble: 155 x 4
#>    title                   year label                      artist          
#>    <chr>                  <int> <chr>                      <chr>           
#>  1 Demo                    2015 Tobi Records (2)           Mollot          
#>  2 Observant Com El Mon …  2013 La Vida Es Un Mus          Una Bèstia Inco…
#>  3 I                       2017 La Vida Es Un Mus          S.H.I.T. (3)    
#>  4 Oído Absoluto           2017 La Vida Es Un Mus          Rata Negra      
#>  5 A Cat's Cause, No Dog…  2015 Katorga Works              Ivy (18)        
#>  6 Tashme                  2019 High Fashion Industries    Tashme          
#>  7 Demo                    2014 Mind Control Records (6)   Desgraciados    
#>  8 Let The Miracles Begin  2015 Not On Label (Phantom Hea… Phantom Head    
#>  9 Sub Space               2017 Not On Label (Sub Space (… Sub Space (2)   
#> 10 Demo                    2017 Prescience Tapes           Small Man (2)   
#> # … with 145 more rows

wesanderson color palettes

wesanderson is the simplest list, containing color palettes, from the wesanderson package. Here’s a glimpse: one component per palette, each containing a character vector of hex colors. Screenshot is of RStudio’s Object Explorer, i.e. from calling View(wesanderson).

library(repurrrsive)
library(purrr)
wesanderson[1:3]
#> $GrandBudapest
#> [1] "#F1BB7B" "#FD6467" "#5B1A18" "#D67236"
#> 
#> $Moonrise1
#> [1] "#F3DF6C" "#CEAB07" "#D5D5D3" "#24281A"
#> 
#> $Royal1
#> [1] "#899DA4" "#C93312" "#FAEFD1" "#DC863B"

Use wesanderson to demonstrate mapping functions over a list.

map_chr(wesanderson, 1)
#>  GrandBudapest      Moonrise1         Royal1      Moonrise2     Cavalcanti 
#>      "#F1BB7B"      "#F3DF6C"      "#899DA4"      "#798E87"      "#D8B70A" 
#>         Royal2 GrandBudapest2      Moonrise3      Chevalier         Zissou 
#>      "#9A8822"      "#E6A0C4"      "#85D4E3"      "#446455"      "#3B9AB2" 
#>   FantasticFox     Darjeeling       Rushmore   BottleRocket    Darjeeling2 
#>      "#DD8D29"      "#FF0000"      "#E1BD6D"      "#A42820"      "#ECCBAE"
map_int(wesanderson, length)
#>  GrandBudapest      Moonrise1         Royal1      Moonrise2     Cavalcanti 
#>              4              4              4              4              5 
#>         Royal2 GrandBudapest2      Moonrise3      Chevalier         Zissou 
#>              5              4              5              4              5 
#>   FantasticFox     Darjeeling       Rushmore   BottleRocket    Darjeeling2 
#>              5              5              5              7              5
map_chr(wesanderson[7:9], paste, collapse = ", ")
#>                                GrandBudapest2 
#>          "#E6A0C4, #C6CDF7, #D8A499, #7294D4" 
#>                                     Moonrise3 
#> "#85D4E3, #F4B5BD, #9C964A, #CDC08C, #FAD77B" 
#>                                     Chevalier 
#>          "#446455, #FDD262, #D3DDDC, #C7B19C"

The same wesanderson data is also present as JSON and XML files. Accessor functions provide the local file path.

wesanderson_json()
#> [1] "/Users/jenny/resources/R/library_3.6/repurrrsive/extdata/wesanderson.json"
wesanderson_xml()
#> [1] "/Users/jenny/resources/R/library_3.6/repurrrsive/extdata/wesanderson.xml"

Practice bringing data from JSON into an R list.

library(jsonlite)
json <- fromJSON(wesanderson_json())
json$wesanderson[1:3]
#> $GrandBudapest
#> [1] "#F1BB7B" "#FD6467" "#5B1A18" "#D67236"
#> 
#> $Moonrise1
#> [1] "#F3DF6C" "#CEAB07" "#D5D5D3" "#24281A"
#> 
#> $Royal1
#> [1] "#899DA4" "#C93312" "#FAEFD1" "#DC863B"
identical(wesanderson, json$wesanderson)
#> [1] TRUE

Practice bringing data into R from XML. You can get it into an R list with xml2::as_list(), but to get a list as nice as those above? That requires a bit more work. Such is XML life.

library(xml2)
xml <- read_xml(wesanderson_xml())
xml_child(xml)
#> {xml_node}
#> <palette name="GrandBudapest">
#> [1] <hex>#F1BB7B</hex>
#> [2] <hex>#FD6467</hex>
#> [3] <hex>#5B1A18</hex>
#> [4] <hex>#D67236</hex>
as_list(xml_child(xml))
#> $hex
#> $hex[[1]]
#> [1] "#F1BB7B"
#> 
#> 
#> $hex
#> $hex[[1]]
#> [1] "#FD6467"
#> 
#> 
#> $hex
#> $hex[[1]]
#> [1] "#5B1A18"
#> 
#> 
#> $hex
#> $hex[[1]]
#> [1] "#D67236"
#> 
#> 
#> attr(,"name")
#> [1] "GrandBudapest"

Nested and split data frame examples

Use the Gapminder data in various forms to practice different styles of grouped computation.

library(dplyr)
library(purrr)
library(tibble)

## group_by() + summarize()
gap_simple %>% 
  group_by(country) %>%
  summarize(cor = cor(lifeExp, year))
#> # A tibble: 142 x 2
#>    country       cor
#>    <fct>       <dbl>
#>  1 Afghanistan 0.974
#>  2 Albania     0.954
#>  3 Algeria     0.993
#>  4 Angola      0.942
#>  5 Argentina   0.998
#>  6 Australia   0.990
#>  7 Austria     0.996
#>  8 Bahrain     0.983
#>  9 Bangladesh  0.995
#> 10 Belgium     0.997
#> # … with 132 more rows

## nest() + map_*() inside mutate()
gap_nested %>%
  mutate(cor = data %>% map_dbl(~ cor(.x$lifeExp, .x$year)))
#> # A tibble: 142 x 4
#>    country     continent data                cor
#>    <fct>       <fct>     <list>            <dbl>
#>  1 Afghanistan Asia      <tibble [12 × 4]> 0.974
#>  2 Albania     Europe    <tibble [12 × 4]> 0.954
#>  3 Algeria     Africa    <tibble [12 × 4]> 0.993
#>  4 Angola      Africa    <tibble [12 × 4]> 0.942
#>  5 Argentina   Americas  <tibble [12 × 4]> 0.998
#>  6 Australia   Oceania   <tibble [12 × 4]> 0.990
#>  7 Austria     Europe    <tibble [12 × 4]> 0.996
#>  8 Bahrain     Asia      <tibble [12 × 4]> 0.983
#>  9 Bangladesh  Asia      <tibble [12 × 4]> 0.995
#> 10 Belgium     Europe    <tibble [12 × 4]> 0.997
#> # … with 132 more rows

## split + map_*()
gap_split %>% 
  map_dbl(~ cor(.x$lifeExp, .x$year)) %>% 
  head()
#> Afghanistan     Albania     Algeria      Angola   Argentina   Australia 
#>   0.9735051   0.9542420   0.9925307   0.9422392   0.9977816   0.9897716

## split + map_*() + tibble::enframe()
gap_split %>% 
  map_dbl(~ cor(.x$lifeExp, .x$year)) %>% 
  enframe()
#> # A tibble: 142 x 2
#>    name        value
#>    <chr>       <dbl>
#>  1 Afghanistan 0.974
#>  2 Albania     0.954
#>  3 Algeria     0.993
#>  4 Angola      0.942
#>  5 Argentina   0.998
#>  6 Australia   0.990
#>  7 Austria     0.996
#>  8 Bahrain     0.983
#>  9 Bangladesh  0.995
#> 10 Belgium     0.997
#> # … with 132 more rows