Tools for functional programming in purrr

Intermediate Functional Programming with purrr

Colin Fay

Data Scientist & R Hacker at ThinkR

High order functions

A high order function can:

Take one or more functions as arguments
Return a function

nop_na <- function(fun){
  function(...){
    fun(..., na.rm = TRUE)
  }
}

sd_no_na <- nop_na(sd)
sd_no_na( c(NA, 1, 2, NA) )

0.7071068

Three types of high order functions

Functionals
Function factories
Function operators

Three types of high order functions in R

Advanced R, Functional Programming

Adverbs in purrr

Handling errors and warnings:

possibly()
safely()

library(purrr)
safe_mean <- safely(mean)
class(safe_mean)

"function"

Use safely() to handle error.

safely() returns a function that will return:

$result
$error

safe_log <- safely(log)
safe_log("a")
$result

NULL

$error

<simpleError in log(x = x, base = base): non-numeric argument to mathematical function>

map( list(2, "a"), log )

Error in log(x = x, base = base) : non-numeric argument to mathematical function

map( list(2, "a"), safely(log) )

[[1]]
[[1]]$result
[1] 0.6931472

[[1]]$error
NULL

[[2]]
[[2]]$result
NULL

[[2]]$error
<simpleError in log(x = x, base = base):non-numeric argument to mathematical function>

Extracting elements from `safely()` results

map() & "result" or "error"

safe_log <- safely(log)

map( list("a", 2),  safe_log) %>% 
  map("result")

[[1]]
NULL

[[2]]
[1] 0.6931472

safe_log <- safely(log)

map( list("a", 2), safe_log ) %>% 
  map("error")

[[1]]
<simpleError in log(x = x, 
base = base): non-numeric argument 
to mathematical function>

[[2]]
NULL

Let's practice!

Intermediate Functional Programming with purrr