Connect operations with the pipe |>

The combination of symbols |> is called the “pipe”. The pipe is a crucial bit of R code.

Why? A bit of backstory first.

We have seen already seen a few examples of R code such as:

# show the dimensions of the data
dim(somedata)

# show a summary of the data
summary(somedata)

# factorise and show the "somevariable" variable in the "somedata" dataframe 
as.factor(somedata$somevariable)

And we can actually wrap functions inside functions:

# factorise the "somevariable" variable in the "somedata" dataframe, 
# then show a summary of it
summary(as.factor(somedata$somevariable))

R evaluates code from the inside-out!

You can end up with functions inside functions inside functions …

# Don't worry about what all these functions do, 
# it's just an example -
round(mean(log(cumsum(diff(1:10)))))
[1] 1

We can write in a different style, however, and this may help to keep code tidy and easily readable - we can write sequentially:

Notice that what we are doing is using a new symbol: |>
This symbol takes the output of whatever is on it’s left-hand side, and uses it as an input for whatever is on the right-hand side. The |> symbol gets called a “pipe”.

Let’s see it in action with the starwars2 dataset. The data contains information on various characteristics of characters from Star Wars. Before we can use the pipe operator, |>, we need to load the tidyverse packages, because that is where |> is found.

library(tidyverse)

starwars2 <- read_csv("https://uoepsy.github.io/data/starwars2.csv")
starwars2 |>
    head()
# A tibble: 6 x 6
  name           height hair_color  eye_color homeworld species
  <chr>           <dbl> <chr>       <chr>     <chr>     <chr>  
1 Luke Skywalker    172 blond       blue      Tatooine  Human  
2 C-3PO             167 <NA>        yellow    Tatooine  Human  
3 R2-D2              96 <NA>        red       Naboo     Droid  
4 Darth Vader       202 none        yellow    Tatooine  Human  
5 Leia Organa       150 brown       brown     Alderaan  Human  
6 Owen Lars         178 brown, grey blue      Tatooine  Human  
starwars2 |>
    summary()
        name        height        hair_color     eye_color      homeworld 
 Length   :75   Min.   : 79   Length   :75   Length   :75   Length   :75  
 N.unique :75   1st Qu.:168   N.unique :12   N.unique :12   N.unique :48  
 N.blank  : 0   Median :180   N.blank  : 0   N.blank  : 0   N.blank  : 0  
 Min.nchar: 5   Mean   :176   Min.nchar: 4   Min.nchar: 3   Min.nchar: 4  
 Max.nchar:21   3rd Qu.:191   Max.nchar:13   Max.nchar:13   Max.nchar:14  
                Max.   :264   NAs      : 6   NAs      : 2                 
      species  
 Length   :75  
 N.unique :37  
 N.blank  : 0  
 Min.nchar: 3  
 Max.nchar:12  
               

We can now write code that requires reading it from the inside-out:

summary(as.factor(starwars2$homeworld))

or which requires reading it from left to right:

starwars2$homeworld |>
    as.factor() |>
    summary()
      Alderaan    Aleen Minor         Bespin     Bestine IV Cato Neimoidia 
             3              1              1              1              1 
         Cerea       Champala      Chandrila   Concord Dawn       Corellia 
             1              1              1              1              2 
     Coruscant       Dathomir          Dorin          Endor         Eriadu 
             3              1              1              1              1 
      Geonosis    Glee Anselm     Haruun Kal        Iktotch       Iridonia 
             1              1              1              1              1 
         Kalee         Kamino       Kashyyyk      Malastare         Mirial 
             1              3              2              1              2 
      Mon Cala     Muunilinst          Naboo      Nal Hutta           Ojom 
             1              1              8              1              1 
       Quermia          Rodia         Ryloth        Serenno          Shili 
             1              1              2              1              1 
         Skako        Socorro    Springfield        Stewjon        Sullust 
             1              1              2              1              1 
      Tatooine       Toydaria      Trandosha        Troiken           Tund 
            10              1              1              1              1 
        Utapau        Vulpter          Zolan 
             1              1              1 

And that long line of code from above:

# again, don't worry about all these functions, 
# just notice the difference in the two styles.
round(mean(log(cumsum(diff(1:10)))))

becomes:

1:10 |>
    diff() |>
    cumsum() |>
    log() |>
    mean() |>
    round()

We’re going to use this way of writing a lot throughout the course, and it pairs really well with a group of functions in the tidyverse packages, which were designed to be used in conjunction with |>.