Functional programming and iteration with purrrMalcolm Barrett2019-09-261 / 68

Interactive app: malcolmbarrett.shinyapps.io/purrr_exercises

Or try `exercises.Rmd` in the repo: github.com/malcolmbarrett/lawrug_purrr

2 / 68

purrr: A functional programming toolkit for R

Complete and consistent set of tools for working with functions and vectors

3 / 68

Problems we want to solve:Making code clear
Making code safe
Working with lists and data frames
4 / 68

Lists, vectors, and data.frames (or tibbles)

c(a = "hello", b = 1)

##       a       b 
## "hello"     "1"

5 / 68

lists can contain any object

list(a = "hello", b = 1, c = mean)

## $a
## [1] "hello"
## 
## $b
## [1] 1
## 
## $c
## function (x, ...) 
## UseMethod("mean")
## <bytecode: 0x7fe91ec2c958>
## <environment: namespace:base>

6 / 68

data frames are lists

x <- list(a = "hello", b = 1)
as.data.frame(x)

##       a b
## 1 hello 1

7 / 68

data frames are lists

library(gapminder)
head(gapminder$pop)

## [1]  8425333  9240934 10267083 11537966 13079460 14880372

8 / 68

data frames are lists

gapminder[1:6, "pop"]

9 / 68

data frames are lists

gapminder[1:6, "pop"]

## # A tibble: 6 x 1
##        pop
##      <int>
## 1  8425333
## 2  9240934
## 3 10267083
## 4 11537966
## 5 13079460
## 6 14880372

10 / 68

data frames are lists

head(gapminder[["pop"]])

## [1]  8425333  9240934 10267083 11537966 13079460 14880372

11 / 68

vectorized functions don't work on lists

sum(rnorm(10))

12 / 68

vectorized functions don't work on lists

sum(rnorm(10))

## [1] -3.831574

13 / 68

vectorized functions don't work on lists

sum(rnorm(10))

## [1] -3.831574

sum(list(x = rnorm(10), y = rnorm(10), z = rnorm(10)))

14 / 68

vectorized functions don't work on lists

sum(rnorm(10))

## [1] -3.831574

sum(list(x = rnorm(10), y = rnorm(10), z = rnorm(10)))

## Error in sum(list(x = rnorm(10), y = rnorm(10), z = rnorm(10))): invalid 'type' (list) of argument

15 / 68

map(.x, .f)16 / 68

map(.x, .f).x: a vector, list, or data frame16 / 68

map(.x, .f).x: a vector, list, or data frame.f: a function16 / 68

map(.x, .f).x: a vector, list, or data frame.f: a functionReturns a list16 / 68

try me!

Using map()

library(purrr)
x_list <- list(x = rnorm(10), y = rnorm(10), z = rnorm(10))
map(x_list, mean)

17 / 68

try me!

Using map()

library(purrr)
x_list <- list(x = rnorm(10), y = rnorm(10), z = rnorm(10))
map(x_list, mean)

18 / 68

try me!

Using map()

library(purrr)
x_list <- list(x = rnorm(10), y = rnorm(10), z = rnorm(10))
map(x_list, mean)

19 / 68

try me!

Using map()

library(purrr)
x_list <- list(x = rnorm(10), y = rnorm(10), z = rnorm(10))
map(x_list, mean)

## $x
## [1] -0.6097971
## 
## $y
## [1] -0.2788647
## 
## $z
## [1] 0.6165922

20 / 68

21 / 68

22 / 68

23 / 68

try me!

using `map()` with data frames

24 / 68

try me!

using `map()` with data frames

gapminder %>% 
  dplyr::select_if(is.numeric) %>% 
  map(sd)

24 / 68

try me!

using `map()` with data frames

gapminder %>%
  dplyr::select_if(is.numeric) %>%
  map(sd)

25 / 68

try me!

using `map()` with data frames

gapminder %>% 
  dplyr::select_if(is.numeric) %>% 
  map(sd)

26 / 68

try me!

using `map()` with data frames

gapminder %>% 
  dplyr::select_if(is.numeric) %>% 
  map(sd)

## $year
## [1] 17.26533
## 
## $lifeExp
## [1] 12.91711
## 
## $pop
## [1] 106157897
## 
## $gdpPercap
## [1] 9857.455

27 / 68

Review: writing functions

x <- x^2
x <- scale(x)
x <- max(x)

28 / 68

Review: writing functions

x <- x^2
x <- scale(x)
x <- max(x)
y <- x^2
y <- scale(y)
y <- max(y)
z <- z^2
z <- scale(x)
z <- max(z)

29 / 68

Review: writing functions

x <- x^2
x <- scale(x)
x <- max(x)
y <- x^2
y <- scale(y)
y <- max(y)
z <- z^2
z <- scale(x)
z <- max(z)

30 / 68

Review: writing functions

x <- x^3
x <- scale(x)
x <- max(x)
y <- x^2
y <- scale(y)
y <- max(y)
z <- z^2
z <- scale(x)
z <- max(z)

31 / 68

Review: writing functions

.f <- function(x) {
  x <- x^3
  x <- scale(x)
  max(x)
}
.f(x)
.f(y)
.f(z)

32 / 68

If you copy and paste your code three times, it's time to write a function33 / 68

Three ways to pass functions to map()pass directly to map()
use an anonymous function
use ~
34 / 68

35 / 68

36 / 68

37 / 68

Annonymous functions

try me!

map(gapminder, ~length(unique(.x)))

38 / 68

Annonymous functions

try me!

## $country
## [1] 142
## 
## $continent
## [1] 5
## 
## $year
## [1] 12
## 
## $lifeExp
## [1] 1626
## 
## $pop
## [1] 1704
## 
## $gdpPercap
## [1] 1704

39 / 68

Returning types

map
returns


map()
list

map_chr()
character vector

map_dbl()
double vector (numeric)

map_int()
integer vector

map_lgl()
logical vector

map_dfc()
data frame (by column)

map_dfr()
data frame (by row)

40 / 68

map	returns
`map()`	list
`map_chr()`	character vector
`map_dbl()`	double vector (numeric)
`map_int()`	integer vector
`map_lgl()`	logical vector
`map_dfc()`	data frame (by column)
`map_dfr()`	data frame (by row)

try me!

Returning types

map_int(gapminder, ~length(unique(.x)))

41 / 68

try me!

Returning types

map_int(gapminder, ~length(unique(.x)))

##   country continent      year   lifeExp       pop gdpPercap 
##       142         5        12      1626      1704      1704

42 / 68

map2(.x, .y, .f)43 / 68

map2(.x, .y, .f).x, .y: a vector, list, or data frame43 / 68

map2(.x, .y, .f).x, .y: a vector, list, or data frame.f: a function43 / 68

map2(.x, .y, .f).x, .y: a vector, list, or data frame.f: a functionReturns a list43 / 68

44 / 68

45 / 68

46 / 68

try me!

map2()

means <- c(-3, 4, 2, 2.3)
sds <- c(.3, 4, 2, 1)
map2_dbl(means, sds, rnorm, n = 1)

47 / 68

try me!

map2()

means <- c(-3, 4, 2, 2.3)
sds <- c(.3, 4, 2, 1)
map2_dbl(means, sds, rnorm, n = 1)

48 / 68

try me!

map2()

means <- c(-3, 4, 2, 2.3)
sds <- c(.3, 4, 2, 1)
map2_dbl(means, sds, rnorm, n = 1)

## [1] -2.997932  2.178125  1.266952  2.948287

49 / 68

input 1
input 2
returns


map()
map2()
list

map_chr()
map2_chr()
character vector

map_dbl()
map2_dbl()
double vector (numeric)

map_int()
map2_int()
integer vector

map_lgl()
map2_lgl()
logical vector

map_dfc()
map2_dfc()
data frame (by column)

map_dfr()
map2_dfr()
data frame (by row)

50 / 68

input 1	input 2	returns
`map()`	`map2()`	list
`map_chr()`	`map2_chr()`	character vector
`map_dbl()`	`map2_dbl()`	double vector (numeric)
`map_int()`	`map2_int()`	integer vector
`map_lgl()`	`map2_lgl()`	logical vector
`map_dfc()`	`map2_dfc()`	data frame (by column)
`map_dfr()`	`map2_dfr()`	data frame (by row)

Other mapping functionspmap() and friends: take n lists or data frame with argument names51 / 68

Other mapping functionspmap() and friends: take n lists or data frame with argument nameswalk() and friends: for side effects like plotting; returns input invisibly52 / 68

Other mapping functionspmap() and friends: take n lists or data frame with argument nameswalk() and friends: for side effects like plotting; returns input invisiblyimap() and friends: includes counter i53 / 68

Other mapping functionspmap() and friends: take n lists or data frame with argument nameswalk() and friends: for side effects like plotting; returns input invisiblyimap() and friends: includes counter imap_if(), map_at(): Apply only to certain elements54 / 68

input 1
input 2
input n
returns


map()
map2()
pmap()
list

map_chr()
map2_chr()
pmap_chr()
character vector

map_dbl()
map2_dbl()
pmap_dbl()
double vector (numeric)

map_int()
map2_int()
pmap_int()
integer vector

map_lgl()
map2_lgl()
pmap_lgl()
logical vector

map_dfc()
map2_dfc()
pmap_dfc()
data frame (by column)

map_dfr()
map2_dfr()
pmap_dfr()
data frame (by row)

walk()
walk2()
pwalk()
input (side effects!)

55 / 68

input 1	input 2	input n	returns
`map()`	`map2()`	`pmap()`	list
`map_chr()`	`map2_chr()`	`pmap_chr()`	character vector
`map_dbl()`	`map2_dbl()`	`pmap_dbl()`	double vector (numeric)
`map_int()`	`map2_int()`	`pmap_int()`	integer vector
`map_lgl()`	`map2_lgl()`	`pmap_lgl()`	logical vector
`map_dfc()`	`map2_dfc()`	`pmap_dfc()`	data frame (by column)
`map_dfr()`	`map2_dfr()`	`pmap_dfr()`	data frame (by row)
`walk()`	`walk2()`	`pwalk()`	input (side effects!)

Base R

base R
purrr


lapply()
map()

vapply()
map_*()

sapply()
?

x[] <- lapply()
map_dfc()

mapply()
map2(), pmap()

56 / 68

base R	purrr
`lapply()`	`map()`
`vapply()`	`map_*()`
`sapply()`	?
`x[] <- lapply()`	`map_dfc()`
`mapply()`	`map2()`, `pmap()`

Benefits of purrrConsistent 
Type-safe
~f(.x)
57 / 68

Efficient Loops

58 / 68

try me!

Loops vs functional programming

x <- rnorm(10)
y <- map(x, mean)

x <- rnorm(10)
y <- vector("list", length(x))
for (i in seq_along(x)) {
  y[[i]] <- mean(x[[i]])
}

59 / 68

try me!

Loops vs functional programming

x <- rnorm(10)
y <- map(x, mean)

x <- rnorm(10) 
y <- vector("list", length(x))
for (i in seq_along(x)) {
  y[[i]] <- mean(x[[i]]) 
}

60 / 68

try me!

Loops vs functional programming

x <- rnorm(10)
y <- map(x, mean)

x <- rnorm(10)
y <- vector("list", length(x)) 
for (i in seq_along(x)) {
  y[[i]] <- mean(x[[i]]) 
}

61 / 68

try me!

Loops vs functional programming

x <- rnorm(10)
y <- map(x, mean)

x <- rnorm(10)
y <- vector("list", length(x)) 
for (i in seq_along(x)) { 
  y[[i]] <- mean(x[[i]])
}

62 / 68

Of course someone has to write loops. It doesn’t have to be you.--Jenny Bryan63 / 68

Working with lists and nested data

64 / 68

Working with lists and nested data

65 / 68

Adverbs: Modify function behavior

66 / 68

Learn more!

Jenny Bryan's purrr tutorial: A detailed introduction to purrr. Free online.

R for Data Science: A comprehensive but friendly introduction to the tidyverse. Free online.

RStudio Primers: Free, interactive courses, including purrr.

67 / 68

Thanks for coming!

map(kittens, wag_tail)

malcolmbarrett

@malco_barrett

Slides created via the R package xaringan.

68 / 68

Help

Keyboard shortcuts

↑, ←, Pg Up, k

Go to previous slide

↓, →, Pg Dn, Space, j

Go to next slide

Home

Go to first slide

End

Go to last slide

Number + Return

Go to specific slide

b / m / f

Toggle blackout / mirrored / fullscreen mode

Clone slideshow

Toggle presenter mode

Restart the presentation timer

?, h

Toggle this help

Functional programming and iteration with purrr

Malcolm Barrett

2019-09-26

Interactive app: malcolmbarrett.shinyapps.io/purrr_exercises

Or try exercises.Rmd in the repo: github.com/malcolmbarrett/lawrug_purrr

purrr: A functional programming toolkit for R

Complete and consistent set of tools for working with functions and vectors

Problems we want to solve:

Lists, vectors, and data.frames (or tibbles)

lists can contain any object

data frames are lists

data frames are lists

data frames are lists

data frames are lists

data frames are lists

vectorized functions don't work on lists

vectorized functions don't work on lists

vectorized functions don't work on lists

vectorized functions don't work on lists

map(.x, .f)

map(.x, .f)

.x: a vector, list, or data frame

map(.x, .f)

.x: a vector, list, or data frame

.f: a function

map(.x, .f)

.x: a vector, list, or data frame

.f: a function

Returns a list

Using map()

Using map()

Using map()

Using map()

using map() with data frames

using map() with data frames

using map() with data frames

using map() with data frames

using map() with data frames

Review: writing functions

Review: writing functions

Review: writing functions

Review: writing functions

Review: writing functions

If you copy and paste your code three times, it's time to write a function

Three ways to pass functions to map()

Annonymous functions

Annonymous functions

Returning types

Returning types

Returning types

map2(.x, .y, .f)

map2(.x, .y, .f)

.x, .y: a vector, list, or data frame

map2(.x, .y, .f)

.x, .y: a vector, list, or data frame

.f: a function

map2(.x, .y, .f)

.x, .y: a vector, list, or data frame

.f: a function

Returns a list

map2()

map2()

map2()

Other mapping functions

pmap() and friends: take n lists or data frame with argument names

Other mapping functions

pmap() and friends: take n lists or data frame with argument names

walk() and friends: for side effects like plotting; returns input invisibly

Other mapping functions

pmap() and friends: take n lists or data frame with argument names

walk() and friends: for side effects like plotting; returns input invisibly

imap() and friends: includes counter i

Other mapping functions

pmap() and friends: take n lists or data frame with argument names

walk() and friends: for side effects like plotting; returns input invisibly

imap() and friends: includes counter i

map_if(), map_at(): Apply only to certain elements

Base R

Benefits of purrr

Efficient Loops

Or try `exercises.Rmd` in the repo: github.com/malcolmbarrett/lawrug_purrr

using `map()` with data frames

using `map()` with data frames

using `map()` with data frames

using `map()` with data frames

using `map()` with data frames

Three ways to pass functions to `map()`

imap() and friends: includes counter `i`

imap() and friends: includes counter `i`

Or try `exercises.Rmd` in the repo: github.com/malcolmbarrett/lawrug_purrr