Unless you need
curry()
orcurry_fn()
, you should use the more versatile gestalt package, which includesfn()
.
Low-Cost Anonymous Functions
nofrills is a lightweight R package that provides fn()
, a more
powerful variation of function()
that:
-
costs less — enables tidyverse quasiquotation so you don’t pay the price of functional impurity
-
has the same great taste — supports a superset of
function()
’s syntax and capabilities -
is less filling —
fn(x, y = 1 ~ x + y)
is equivalent to
function(x, y = 1) x + y
install.packages("nofrills")
Alternatively, install the development version from GitHub:
# install.packages("devtools")
devtools::install_github("egnha/nofrills")
fn(x ~ x + 1)
#> function (x)
#> x + 1
fn(x, y ~ x + y)
#> function (x, y)
#> x + y
fn(x, y = 2 ~ x + y)
#> function (x, y = 2)
#> x + y
fn(x, y = 1, ... ~ log(x + y, ...))
#> function (x, y = 1, ...)
#> log(x + y, ...)
# the only exception, cf. alist()
fn(x, ... = , y ~ log(x + y, ...))
#> function (x, ..., y)
#> log(x + y, ...)
fn(~ NA)
#> function ()
#> NA
z <- 0
fn(x, y = !!z ~ x + y)
#> function (x, y = 0)
#> x + y
fn(x ~ x > !!z)
#> function (x)
#> x > 0
arg <- "y"
fn(x, !!arg := 0 ~ x + !!as.name(arg))
#> function (x, y = 0)
#> x + y
args <- alist(x, y = 0)
fn(!!!args, ~ x + y) # note the one-sided formula
#> function (x, y = 0)
#> x + y
library(dplyr)
summariser <- quote(mean)
my_summarise <- fn(df, ... ~ {
group_by <- quos(...)
df %>%
group_by(QUQS(group_by)) %>%
summarise(a = (!!summariser)(a))
})
my_summarise
#> function (df, ...)
#> {
#> group_by <- quos(...)
#> df %>% group_by(`!!!`(group_by)) %>% summarise(a = mean(a))
#> }
(Source: Programming with dplyr)
Curry functions
The syntax is the same as fn()
. Using the literal unquoting operators
QUQ()
, QUQS()
, you can “delay” unquoting to embed argument values in
the innermost function:
compare_to <- curry_fn(target, x ~ identical(x, QUQ(target)))
is_this <- compare_to("this")
# The embedded value "this" renders the source comprehensible
is_this
#> function (x)
#> identical(x, "this")
#> <environment: 0x7fdc55943678>
curry(function(x, y, z = 0) x + y + z)
#> function (x)
#> function(y) function(z = 0) x + y + z
double <- curry(`*`)(2)
double(3)
#> [1] 6
Functions in R are generally impure, i.e., the return value of a function will not in general be determined by the value of its inputs alone. This is because a function may depend on mutable objects in its lexical scope. Normally this isn’t an issue. But if you are working interactively and sourcing files into the global environment, say, or using a notebook interface (like Jupyter or R Notebook), it can be tricky to ensure that you haven’t unwittingly mutated an object that an earlier function depends upon.
-
Consider the following function:
a <- 1 foo <- function(x) x + a
What is the value of
foo(1)
? It is not necessarily2
because the value ofa
may have changed between the creation offoo()
and the calling offoo(1)
:foo(1) #> [1] 2 a <- 0 foo(1) #> [1] 1
In other words,
foo()
is impure because the value offoo(x)
depends not only on the value ofx
but also on the externally mutable value ofa
.
fn()
enables you to write pure(r) functions by using
quasiquotation to eliminate such indeterminacy.
-
With
fn()
, you can unquotea
to capture its value at the point of creation:a <- 1 foo <- fn(x ~ x + !!a)
Now
foo()
is a pure function, unaffected by changes in its lexical scope:foo(1) #> [1] 2 a <- 0 foo(1) #> [1] 2
Alternative anonymous-function constructors (which don’t support quasiquotation) include:
The rlang package by Lionel Henry and Hadley Wickham makes nofrills possible. Crucially, rlang provides the engine for quasiquotation and expression capture.
MIT Copyright © 2017–22 Eugene Ha