4 Pipes

4.1 Introduction

Use |> to emphasise a sequence of actions, rather than the object that the actions are being performed on.

The tidyverse has been designed to work particularly well with the pipe, but you can use it with any code, particularly in conjunction with the _ placeholder.

strings |>
  str_replace("a", "b") |>
  str_replace("x", "y")

strings |>
  gsub("a", "b", x = _) |>
  gsub("x", "y", x = _)

Avoid using the pipe when:

You need to manipulate more than one object at a time. Reserve pipes for a sequence of steps applied to one primary object.
There are meaningful intermediate objects that could be given informative names.

4.2 Whitespace

|> should always have a space before it, and should usually be followed by a new line. After the first step, each line should be indented by two spaces. This structure makes it easier to add new steps (or rearrange existing steps) and harder to overlook a step.

# Good
iris |>
  summarize(across(where(is.numeric), mean), .by = Species) |>
  pivot_longer(!Species, names_to = "measure", values_to = "value") |>
  arrange(value)

# Bad
iris |> summarize(across(where(is.numeric), mean), .by = Species) |>
pivot_longer(!Species, names_to = "measure", values_to = "value") |>
arrange(value)

4.3 Long lines

If the arguments to a function don’t all fit on one line, put each argument on its own line and indent:

# Good
iris |>
  summarise(
    Sepal.Length = mean(Sepal.Length),
    Sepal.Width = mean(Sepal.Width),
    .by = Species
  )

# Bad
iris |>
  summarise(Sepal.Length = mean(Sepal.Length), Sepal.Width = mean(Sepal.Width), .by = Species)

For data analysis, we recommend using the pipe whenever a function needs to span multiple lines, even if it’s only a single step.

# Bad
summarise(
  iris,
  Sepal.Length = mean(Sepal.Length),
  Sepal.Width = mean(Sepal.Width),
  .by = Species
)

4.4 Short pipes

It’s ok to write a short pipe on a single line:

# Ok
iris |> subset(Species == "virginica") |> _$Sepal.Length
iris |> summarise(width = Sepal.Width, .by = Species) |> arrange(width)

But because short pipes often become longer pipes, we recommend that you generally stick to one function per line:

# Better
iris |>
  subset(Species == "virginica") |>
  _$Sepal.Length

iris |>
  summarise(width = Sepal.Width, .by = Species) |>
  arrange(width)

Sometimes it’s useful to include a short pipe as an argument to a function in a longer pipe. Carefully consider whether the code is more readable with a short inline pipe (which doesn’t require a lookup elsewhere) or if it’s better to move the code outside the pipe and give it an evocative name.

# Good
x |>
  semi_join(y |> filter(is_valid))

# Ok
x |>
  select(a, b, w) |>
  left_join(y |> select(a, b, v), join_by(a, b))

# Better
x_join <- x |> select(a, b, w)
y_join <- y |> select(a, b, v)
left_join(x_join, y_join, join_by(a, b))

4.5 Assignment

There are three acceptable forms of assignment:

Variable name and assignment on separate lines:

iris_long <-
  iris |>
  gather(measure, value, -Species) |>
  arrange(-value)

Variable name and assignment on the same line:

iris_long <- iris |>
  gather(measure, value, -Species) |>
  arrange(-value)

Assignment at the end of the pipe with ->:

iris |>
  gather(measure, value, -Species) |>
  arrange(-value) ->
  iris_long

I think that the third is the most natural to write, but makes reading a little harder: when the name comes first, it can act as a heading to remind you of the purpose of the pipe.

4.6 magrittr

We recommend you use the base |> pipe instead of magrittr’s %>%.

# Good
iris |>
  summarise(width = Sepal.Width, .by = Species) |>
  arrange(width)

# Bad
iris %>%
  summarise(width = Sepal.Width, .by = Species) %>%
  arrange(width)

As of R 4.3.0, the base pipe provides all the features from magrittr that we recommend using.