Penguins and nakedpipes

Exploring the new {palmerpenguins} dataset with {nakedpipe} - An alternative to using {magrittr}’s %>%.

Published

June 21, 2020

Recently I came across nakedpipe Fabri (2020) on github - an R package with an intriguing name to say the least… After checking out the repo, I was very impressed and thought I’d share my notes here!

So what does nakedpipe do?

I’m sure most R users are familiar with magrittr’s pipe (%>%) operator. The %>% allows you to chain together multiple commands by piping forward an expression into a function like x %>% f, rather than f(x) Bache and Wickham (2014). The resulting code is usually very readable and easy to debug. nakedpipe adopts a similar style, but removes the need to use %>% after every function, and adds some additional logging/debugging features as well as being slightly faster than the magrittr implementation.

In order to try out the features in nakedpipe I used the new palmerpenguins dataset (KB, TD, and WR 2014) which provides a great test dataset and is a refreshing alternative the usual suspects, mtcars et al…

Getting started

#remotes::install_github("allisonhorst/palmerpenguins")
#remotes::install_github("moodymudskipper/nakedpipe")

library(tidyverse)
library(palmerpenguins) 
library(nakedpipe)
library(magrittr)

penguins <- penguins

A basic example looks like this:

penguins %.% {
  filter(sex == "female") 
  select(1:5) 
  head(5)
}

# A tibble: 5 × 5
  species island    bill_length_mm bill_depth_mm flipper_length_mm
  <fct>   <fct>              <dbl>         <dbl>             <int>
1 Adelie  Torgersen           39.5          17.4               186
2 Adelie  Torgersen           40.3          18                 195
3 Adelie  Torgersen           36.7          19.3               193
4 Adelie  Torgersen           38.9          17.8               181
5 Adelie  Torgersen           41.1          17.6               182

You use the %.% operator to “pipe” into a sequence of functions, which are within {}. You can reduce the code further by removing the subset / filter argument like:

Any call to < > <= >= == != %in% & | will work in this way

penguins %.% {
  sex == "female"
  select(1:5) 
  head(5)
}

# A tibble: 5 × 5
  species island    bill_length_mm bill_depth_mm flipper_length_mm
  <fct>   <fct>              <dbl>         <dbl>             <int>
1 Adelie  Torgersen           39.5          17.4               186
2 Adelie  Torgersen           40.3          18                 195
3 Adelie  Torgersen           36.7          19.3               193
4 Adelie  Torgersen           38.9          17.8               181
5 Adelie  Torgersen           41.1          17.6               182

Similarly, this works with transform / mutate calls. The use of the = sign implies a new column is being created, for example:

penguins %.% {
  sex == "female"
  body_mass_kg = body_mass_g/1000
  select(body_mass_g, body_mass_kg)
  head(5)
}

  body_mass_g body_mass_kg
1        3800        3.800
2        3250        3.250
3        3450        3.450
4        3625        3.625
5        3200        3.200

Assigning the result of a sequence of function can be done by using -> at the end of the sequence (outside of the {}) or using the %<.% operator at the start of the sequence.

penguins %.% {
  sex == "female"
  arrange(body_mass_g)
  select(1,2,6) 
  head(5)
} -> small_penguins

# Or

small_penguins <- penguins

small_penguins %<.% {
  sex == "female"
  arrange(body_mass_g)
  select(1,2,6) 
  head(5)
}

Additionally, you can create outputs midway through the sequence of functions by using ~~:

penguins %.% {
  sex == "female"
  ~~ . -> female_penguins
  select(1:5) 
  head(5) 
} -> x

head(female_penguins)

# A tibble: 6 × 8
  species island bill_length_mm bill_depth_mm flipper_length_… body_mass_g sex  
  <fct>   <fct>           <dbl>         <dbl>            <int>       <int> <fct>
1 Adelie  Torge…           39.5          17.4              186        3800 fema…
2 Adelie  Torge…           40.3          18                195        3250 fema…
3 Adelie  Torge…           36.7          19.3              193        3450 fema…
4 Adelie  Torge…           38.9          17.8              181        3625 fema…
5 Adelie  Torge…           41.1          17.6              182        3200 fema…
6 Adelie  Torge…           36.6          17.8              185        3700 fema…
# … with 1 more variable: year <int>

Additional operators

You can use %P.% to print the output of each step:

penguins %P.% {
  sex == "female"
  select(1:5) 
  head(5)
} -> y

penguins %P.% {

  sex == "female"

# A tibble: 165 × 8
   species island    bill_length_mm bill_depth_mm flipper_length_mm body_mass_g
   <fct>   <fct>              <dbl>         <dbl>             <int>       <int>
 1 Adelie  Torgersen           39.5          17.4               186        3800
 2 Adelie  Torgersen           40.3          18                 195        3250
 3 Adelie  Torgersen           36.7          19.3               193        3450
 4 Adelie  Torgersen           38.9          17.8               181        3625
 5 Adelie  Torgersen           41.1          17.6               182        3200
 6 Adelie  Torgersen           36.6          17.8               185        3700
 7 Adelie  Torgersen           38.7          19                 195        3450
 8 Adelie  Torgersen           34.4          18.4               184        3325
 9 Adelie  Biscoe              37.8          18.3               174        3400
10 Adelie  Biscoe              35.9          19.2               189        3800
# … with 155 more rows, and 2 more variables: sex <fct>, year <int>

  select(1:5)

# A tibble: 165 × 5
   species island    bill_length_mm bill_depth_mm flipper_length_mm
   <fct>   <fct>              <dbl>         <dbl>             <int>
 1 Adelie  Torgersen           39.5          17.4               186
 2 Adelie  Torgersen           40.3          18                 195
 3 Adelie  Torgersen           36.7          19.3               193
 4 Adelie  Torgersen           38.9          17.8               181
 5 Adelie  Torgersen           41.1          17.6               182
 6 Adelie  Torgersen           36.6          17.8               185
 7 Adelie  Torgersen           38.7          19                 195
 8 Adelie  Torgersen           34.4          18.4               184
 9 Adelie  Biscoe              37.8          18.3               174
10 Adelie  Biscoe              35.9          19.2               189
# … with 155 more rows

  head(5)

# A tibble: 5 × 5
  species island    bill_length_mm bill_depth_mm flipper_length_mm
  <fct>   <fct>              <dbl>         <dbl>             <int>
1 Adelie  Torgersen           39.5          17.4               186
2 Adelie  Torgersen           40.3          18                 195
3 Adelie  Torgersen           36.7          19.3               193
4 Adelie  Torgersen           38.9          17.8               181
5 Adelie  Torgersen           41.1          17.6               182

%L.% prints out timings of each step in the sequence (Not particularly useful for this toy example, but would be for longer running code)

penguins %L.% {
  sex == "female"
  select(1:5) 
  head(5)
} -> z

penguins %L.% {

  sex == "female"

   user  system elapsed 
  0.001   0.000   0.001

  select(1:5)

   user  system elapsed 
  0.009   0.000   0.009

  head(5)

   user  system elapsed 
  0.001   0.000   0.000

%F.% lets you assign a function using the nakedpipe syntax:

penguin_func <- . %F.% {
  group_by(species)
  summarise(across(where(is.numeric), ~mean(., na.rm = TRUE)))
  mutate(across(where(is.numeric), round, 2))
}

penguin_func(penguins)

# A tibble: 3 × 6
  species   bill_length_mm bill_depth_mm flipper_length_mm body_mass_g  year
  <fct>              <dbl>         <dbl>             <dbl>       <dbl> <dbl>
1 Adelie              38.8          18.4              190.       3701. 2008.
2 Chinstrap           48.8          18.4              196.       3733. 2008.
3 Gentoo              47.5          15.0              217.       5076. 2008.

ggplot

nakedpipe also plays well with ggplot:

penguins %.% {
  body_mass_kg = body_mass_g / 1000
  ggplot(aes(body_mass_kg, flipper_length_mm, shape = species, colour = species))
  + geom_point() 
  + labs(y = "Flipper length", x = "Body mass (kg)")
  + theme_minimal()
}

There are some other useful things that nakepipe can do that I haven’t included here, such as %D.% operator for debugging, %..% for even faster performance and even a RStudio addin to convert your magrittr style code to the nakepipe syntax. Pretty cool!

If you’ve found any of these examples interesting, I’d highly recommend you check out the package (https://github.com/moodymudskipper/nakedpipe) and give it a try!

References

Bache, Stefan Milton, and Hadley Wickham. 2014. Magrittr: A Forward-Pipe Operator for r. https://CRAN.R-project.org/package=magrittr.

Fabri, Antoine. 2020. Nakedpipe: Pipe into a Sequence of Calls Without Repeating the Pipe Symbol. https://github.com/moodymudskipper/nakedpipe.

KB, Gorman, Williams TD, and Fraser WR. 2014. “Ecological Sexual Dimorphism and Environmental Variability Within a Community of Antarctic Penguins (Genus Pygoscelis).” PLoS ONE 9(3) (e90081): –13. https://doi.org/10.1371/journal.pone.0090081.