Skip to contents

Parallelize 'purrr' functions

Henrik Bengtsson

Source: vignettes/futurize-21-purrr.md
futurize-21-purrr.Rmd

The 'purrr' logo+ The 'futurize' hexlogo= The 'future' logo

The futurize package allows you to easily turn sequential code into parallel code by piping the sequential code to the futurize() function. Easy!

TL;DR 

library(futurize)
plan(multisession)
library(purrr)

slow_fcn <- function(x) {
  Sys.sleep(0.1)  # emulate work
  x^2
}

xs <- 1:1000
ys <- xs |> map(slow_fcn) |> futurize()

Introduction

This vignette demonstrates how to use this approach to parallelize purrr functions such as map(), map_dbl(), and walk().

The purrr map() function is commonly used to apply a function to the elements of a vector or a list. For example,

library(purrr)
xs <- 1:1000
ys <- map(xs, slow_fcn)

or equivalently using pipe syntax

library(purrr)
xs <- 1:1000
ys <- xs |> map(slow_fcn)

Here map() evaluates sequentially, but we can easily make it evaluate in parallel, by using:

library(futurize)
library(purrr)
xs <- 1:1000
ys <- xs |> map(slow_fcn) |> futurize()

This will distribute the calculations across the available parallel workers, given that we have set parallel workers, e.g.

plan(multisession)

The built-in multisession backend parallelizes on your local computer and it works on all operating systems. There are other parallel backends to choose from, including alternatives to parallelize locally as well as distributed across remote machines, e.g.

plan(future.mirai::mirai_multisession)

and

plan(future.batchtools::batchtools_slurm)

Another example is:

library(purrr)
library(futurize)
plan(future.mirai::mirai_multisession)

ys <- 1:10 |>
        map(rnorm, n = 10) |> futurize(seed = TRUE) |>
        map_dbl(mean) |> futurize()

Supported Functions

The futurize() function supports parallelization of the following purrr functions: