pipe

This package proposes an alternative tot he pipe from ‘magrittr’. It’s named the same and passes all the ‘magrittr’ test so can easily be a drop-in replacement, its main advantages is that it is implemented more consistently and allows definition of new pipes. It also solves a few issues of the original pipe.

Install with :

remotes::install_github("pipe")

What’s different from magrittr ?

The main differences are:

Pipes are object of class pipe and of class either standard_pipe or compound_pipe, none of them is special implementation-wise
Every step of the chain is executed in the same environment, as can be seen by the use of %B>% above
The new pipe %B>% allows browsing the pipe chain step by step

calling iris %>% head %B>% dim %>% length will open place you right at the browser call below :

#> Called from: (function (.) 
#> {
#>     on.exit(rm(.))
#>     . <- head(.)
#>     browser()
#>     . <- dim(.)
#>     . <- length(.)
#>     .
#> })(iris)

The new pipe %S>% allows the use of rlang’s !!! operator to splice dots in any function

library(pipe)
c(a = 1, b = 2) %S>% data.frame()
#>   .
#> a 1
#> b 2

additional pipes can be defined easily by calling new_pipe(), which is used internally to create pipes included in the package
:::, :: and $ get a special treatment to solve a common issue

iris %>% base::dim
#> [1] 150   5
iris %>% base:::dim
#> [1] 150   5
x <- list(y = dim)
iris %>% x$y
#> [1] 150   5

It fails explicitly in some cases rather than allowing strange behavior silently

iris %>% head %<>% dim
#> Error: A compound pipe should only be used at the start of the chain
. %<>% head
#> Error in . %<>% head: You can't start a functional sequence on a compound operator

It’s slightly slower than magrittr (see fast_pipe if speed is an issue!)

microbenchmark::microbenchmark(
  magrittr = {`%>%` <- magrittr::`%>%`
  1 %>% identity %>% identity() %>% (identity) %>% {identity(.)}},
  pipe = {`%>%` <- pipe::`%>%`
  1 %>% identity %>% identity() %>% (identity) %>% {identity(.)}},
  times=10000
)
#> Unit: microseconds
#>      expr   min     lq     mean median    uq     max neval cld
#>  magrittr  93.0 118.00 215.5137  152.3 240.9  5113.1 10000  a 
#>      pipe 128.9 157.15 289.4841  204.4 330.4 11374.4 10000   b

Defining new pipes

Pipes are defined by the function new_pipe() which accepts arguments format_fun and compound_fun.

format_fun is the function that changes the quoted rhs of a pipe into the quoted expression that will give a new value to the dot. For instance in iris %>% head() its input will be quote(head()) and it’s output quote(. <- head(.)).

compound_fun is used to define new types of compound pipes, we won’t detail it as it’s very specific, but see last example, used to define %<>%.

We enumerate all the pipes of the package below :

`%>%` <- new_pipe(
  function(call){
  call <- insert_dot(call)
  bquote(. <- .(call))
})

`%T>%`<- new_pipe(
  function(call){
  call <- insert_dot(call)
  bquote(local(.(call)))
})

`%$%`<- new_pipe(
  function(call){
  bquote(. <- with(.,.(call)))
})

`%S>%`<- new_pipe(
  function(call){
    call <- insert_dot(call)
    bquote(. <- eval(rlang::expr(.(call))))
  })


`%B>%` <- new_pipe(
  function(call){
  call <- insert_dot(call)
  c(quote(browser()),
    bquote(. <- .(call)))
})

`%<>%` <- new_pipe(
  function(call){
    call <- insert_dot(call)
    bquote(. <- .(call))
  },
  function(lhs, res){
    substitute(lhs <- res)
  })

A few notes

This is just an experiment, it’s likely to change. I’ve also been working on the package fastpipe which is a closer equivalent to magrittr focused on being more robust and faster.

trinker/pipe

pipe

What’s different from magrittr ?

Defining new pipes

A few notes