Substitute the magrittr Pipe %>% with R’s Native Pipe Operator |>

Yihui Xie 2022-04-28

R 4.1.0 has added a native pipe operator |> to base R. If you are using the magrittr pipe %>% and want to switch to |> in your R scripts (you do not have to), I have a quick function formatR::tidy_pipe() that can help you substitute %>% with |> in the RStudio editor. All you need to do is open the R script in RStudio, and run formatR::tidy_pipe() in the R console. Then you will see %>%’s are converted to |>:

'%>%' %>%
  c('->', '|>') %>%
  paste(collapse = ' ')

This function only works with the R script currently opened in the RStudio editor, but it is definitely possible to remove this limitation (see the end of this post).

You can also use the function formatR::tidy_source() to reformat code, including substituting %>% with |>, e.g.,

formatR::tidy_source('your-code.R', pipe = TRUE)

Note that this function does a lot other reformatting, too. If you are not familiar with it, you may take a look at its documentation.

P.S. The word “tidy” in formatR has nothing to do with Tidyverse. It came from 2007 and means reformatting.

An exercise

If you are interested in helping with extending formatR::tidy_pipe(), below is the basic idea about how it works under the hood:

  1. Use getParseData(parse(text = code)) to get the data about the code.

    > formatR:::parse_data(
    "'%>%' %>% c('->', '|>') %>% paste(collapse = ' ')"
  )

   line1 col1 line2 col2 id                 token      text
1      1    1     1    5  1             STR_CONST     '%>%'
2      1    7     1    9  2               SPECIAL       %>%
4      1   11     1   11  4  SYMBOL_FUNCTION_CALL         c
5      1   12     1   12  5                   '('         (
....

  2. Find the symbols %>% in the data, and their row/column indices.

  3. Substitute the %>% strings in the original code with |> using the indices.

The 3rd step can be a little tricky if you use substr<- but I have an idea to make it work. Please feel free to open a pull request on GitHub if you are interested in contributing (it is okay if you are unable to figure out everything—we can talk on GitHub). Thanks!

Endless pipes