7.9 The Native Pipe Operator

In Section 7.1, we introduced the pipe operator %>% from the magrittr package (loaded automatically with the tidyverse). Starting from R version 4.1.0 (released in 2021), R includes a built-in native pipe operator |> that works without loading any packages.

7.9.1 Syntax Comparison

Both pipe operators take the result of the expression on the left and pass it as the first argument to the function on the right.

library(r02pro)
library(dplyr)

# Using the magrittr pipe
sahp %>%
  filter(house_style == "2Story") %>%
  summarize(avg_price = mean(sale_price))
#> # A tibble: 1 × 1
#>   avg_price
#>       <dbl>
#> 1        NA

# Using the native pipe
sahp |>
  filter(house_style == "2Story") |>
  summarize(avg_price = mean(sale_price))
#> # A tibble: 1 × 1
#>   avg_price
#>       <dbl>
#> 1        NA

As you can see, they produce exactly the same result.

7.9.2 Key Differences

While the two pipe operators behave similarly in most cases, there are a few important differences:

  1. No placeholder: The %>% pipe supports . as a placeholder to insert the left-hand side into a position other than the first argument. The native pipe |> does not support a placeholder in the same way. Instead, you can use an anonymous function.
# magrittr pipe with placeholder
c(3, 1, 2) %>% paste("Item", .)
#> [1] "Item 3" "Item 1" "Item 2"

# native pipe with anonymous function
c(3, 1, 2) |> (\(x) paste("Item", x))()
#> [1] "Item 3" "Item 1" "Item 2"

  1. No package needed: The native pipe |> is built into R, so you don’t need to load any package to use it.

  2. Speed: The native pipe is slightly faster since it is implemented at the language level rather than as a function call.

7.9.3 Which One Should You Use?

Both pipes are widely used in the R community. Here are some guidelines:

  • If you are writing a tidyverse-heavy script and already loading dplyr or magrittr, either pipe works well.
  • If you want to minimize package dependencies, use the native pipe |>.
  • If you need the placeholder feature frequently, %>% may be more convenient.
  • In this book, we primarily use %>% for consistency with existing examples, but you are encouraged to try |> in your own work.

7.9.4 Exercises

  1. Rewrite the following code using the native pipe |> instead of %>%:
sahp %>%
  filter(oa_qual > 5) %>%
  select(sale_price, liv_area) %>%
  arrange(desc(sale_price)) %>%
  head(10)

  1. Using the gm2004 dataset, write a pipeline with the native pipe |> that filters for countries in "Europe", selects country and life_expectancy, and arranges by life_expectancy in descending order.

  2. Explain why the following code using the native pipe does not work, and rewrite it so that it does:

c(1, 5, 3, 2, 4) |> sort() |> paste("Value:", .)

Buy Me A Coffee