Functions and Iterations

Learning to do repeated tasks in R

Meenakshi Kushwaha

18th August, 2022

Functions

When should you write function?

You should consider writing a function whenever you’ve copied and pasted a block of code more than twice (i.e. you now have three copies of the same code).”
- R for Data Science

Example of a repeated task

In the gapminder dataset calculate the range of population, gdp, and life expectancy

round(max(gapminder$lifeExp) - min(gapminder$lifeExp),1)
[1] 59
round(max(gapminder$pop) - min(gapminder$pop),1)
[1] 1318623085
round(max(gapminder$gdpPercap) - min(gapminder$gdpPercap),1)
[1] 113282

Your first function

max_minus_min <- function(x){
  round((max(x) - min(x)), 1)
  }

Test-run your function

max_minus_min(gapminder$lifeExp)
[1] 59

What did we need?

  • name of the function

max_minus_min

  • what do you want the function to do? OR body

round((max(x) - min(x)), 1)

  • what does the function need to work? OR ingredient

x

max_minus_min <- function(x){
  round((max(x) - min(x)), 1)
  }

A function returns the last statement it evaluates

Quiz

What will the following code return?

addTen <- function(x){
  x+10
}

addTen(20)
  • 10
  • 20
  • 30
  • Will give an error

How to re-use your function

  • Step 1: Save your function as .R script
  • Step 2: Source the script using
source("your-function-name.R")

This will work if your .R script is in your working directory, else specify the path

source("path/tofile/your-function-name.R")

Demo

Good Practices

  • descriptive names, use verbs for function names
  • use snake case in names
  • be consistent with naming
  • use common prefix for a family of related functions

Other tips

  • Use comments for #why. Code should explan the #how
  • Shorcut (extract function)

Iterations

map() Function

FOR EACH ____ DO _____

map(.x, .f)

.x is a vector/list/dataframe
.f is the action that you want to do with each element of .x

Types of map() functions

  • map(.x, .f) is the main mapping function and returns a list

  • map_df(.x, .f) returns a data frame

  • map_dbl(.x, .f) returns a numeric (double) vector

  • map_chr(.x, .f) returns a character vector

  • map_lgl(.x, .f) returns a logical vector

Input to map() functions

  • a vector (of any type)
    • in which case the iteration is done over the entries of the vector
  • a list
    • in which case the iteration is performed over the elements of the list
  • a data frame
    • in which case the iteration is performed over the columns of the data frame

Example

Apply class() function to each column of gampminder data

gapminder %>% 
  map(class)
$country
[1] "factor"

$continent
[1] "factor"

$year
[1] "integer"

$lifeExp
[1] "numeric"

$pop
[1] "integer"

$gdpPercap
[1] "numeric"
gapminder %>% 
  map_chr(class)
  country continent      year   lifeExp       pop gdpPercap 
 "factor"  "factor" "integer" "numeric" "integer" "numeric"

Example

What is the number of distinct values in each column

gapminder %>% 
  map_int(n_distinct)
  country continent      year   lifeExp       pop gdpPercap 
      142         5        12      1626      1704      1704

Example

What is the median of all numeric columns?

gapminder %>% 
  dplyr::select_if(is.numeric) %>% 
  map_dbl(median)
        year      lifeExp          pop    gdpPercap 
   1979.5000      60.7125 7023595.5000    3531.8470

When things get complicated, use ~

Sometimes you may want to define your own function and apply to each column

function(x) { x + 10 }

can be replaced with ~{. + 10}

Using ~ with map functions

my_vector <- c(1, 2, 3)
map_dbl(my_vector, function(x){x+10})

is same as

my_vector <- c(1, 2, 3)
map_dbl(my_vector, ~(.+10))

Quiz

What is the output of the following code?

my_vector <- c(1, 2, 3)
map_dbl(my_vector, ~(.+10))
my_vector <- c(1, 2, 3)
map_dbl(my_vector, ~(.+10))
[1] 11 12 13

~(.+10) adds 10 to each element of my_vector

Using ~ with map functions

map_int(gapminder, ~length(unique(.)))
  country continent      year   lifeExp       pop gdpPercap 
      142         5        12      1626      1704      1704

Using ~ with map functions

Fitting a linear model with different groups of the data

gapminder %>% 
  split(.$continent) %>% 
  map(function(df) lm(lifeExp ~ pop, data = df))
$Africa

Call:
lm(formula = lifeExp ~ pop, data = df)

Coefficients:
(Intercept)          pop  
  4.816e+01    7.150e-08  


$Americas

Call:
lm(formula = lifeExp ~ pop, data = df)

Coefficients:
(Intercept)          pop  
  6.353e+01    4.587e-08  


$Asia

Call:
lm(formula = lifeExp ~ pop, data = df)

Coefficients:
(Intercept)          pop  
  5.992e+01    1.901e-09  


$Europe

Call:
lm(formula = lifeExp ~ pop, data = df)

Coefficients:
(Intercept)          pop  
  7.162e+01    1.650e-08  


$Oceania

Call:
lm(formula = lifeExp ~ pop, data = df)

Coefficients:
(Intercept)          pop  
  7.207e+01    2.545e-07

Using ~ with map functions

Fitting a linear model with different groups of the data

gapminder %>% 
  split(.$continent) %>% 
  map(~lm(lifeExp ~ pop, data = .))
$Africa

Call:
lm(formula = lifeExp ~ pop, data = .)

Coefficients:
(Intercept)          pop  
  4.816e+01    7.150e-08  


$Americas

Call:
lm(formula = lifeExp ~ pop, data = .)

Coefficients:
(Intercept)          pop  
  6.353e+01    4.587e-08  


$Asia

Call:
lm(formula = lifeExp ~ pop, data = .)

Coefficients:
(Intercept)          pop  
  5.992e+01    1.901e-09  


$Europe

Call:
lm(formula = lifeExp ~ pop, data = .)

Coefficients:
(Intercept)          pop  
  7.162e+01    1.650e-08  


$Oceania

Call:
lm(formula = lifeExp ~ pop, data = .)

Coefficients:
(Intercept)          pop  
  7.207e+01    2.545e-07

Strategy

  • Solve for one element
  • Turn it into a recipe
  • Use map() to solve for all elements

Demo

Reading multiple files with purrr package

Reading multiple files

We have data from different countries in seperate .csv files that we need to combine in a single dataset for analysis

Step 1

Find all .csv files in a folder

dir(here("data"), pattern = "*.csv") #list all .csv files in data folder
[1] "china_gm.csv" "india_gm.csv" "japan_gm.csv" "nepal_gm.csv"

Reading multiple files

Step 2

Make a list of all the files

my_files <- dir(here("data"), # specify file path
                pattern = "*.csv",  # look for .csv files
                full.names = TRUE)  # preserve file path
my_files
[1] "/Users/meenakshikushwaha/Dropbox/R projects/github/cstep-iterations/data/china_gm.csv"
[2] "/Users/meenakshikushwaha/Dropbox/R projects/github/cstep-iterations/data/india_gm.csv"
[3] "/Users/meenakshikushwaha/Dropbox/R projects/github/cstep-iterations/data/japan_gm.csv"
[4] "/Users/meenakshikushwaha/Dropbox/R projects/github/cstep-iterations/data/nepal_gm.csv"

Reading multiple files

Step 3

Read and combine all files using map_dfr()

my_files %>% 
  map_dfr(read_csv) 
# A tibble: 16 × 6
   country continent  year lifeExp       pop gdpPercap
   <chr>   <chr>     <dbl>   <dbl>     <dbl>     <dbl>
 1 China   Asia       1952    44   556263527      400.
 2 China   Asia       1957    50.5 637408000      576.
 3 China   Asia       1962    44.5 665770000      488.
 4 China   Asia       1967    58.4 754550000      613.
 5 India   Asia       1952    37.4 372000000      547.
 6 India   Asia       1957    40.2 409000000      590.
 7 India   Asia       1962    43.6 454000000      658.
 8 India   Asia       1967    47.2 506000000      701.
 9 Japan   Asia       1952    63.0  86459025     3217.
10 Japan   Asia       1957    65.5  91563009     4318.
11 Japan   Asia       1962    68.7  95831757     6577.
12 Japan   Asia       1967    71.4 100825279     9848.
13 Nepal   Asia       1952    36.2   9182536      546.
14 Nepal   Asia       1957    37.7   9682338      598.
15 Nepal   Asia       1962    39.4  10332057      652.
16 Nepal   Asia       1967    41.5  11261690      676.

Using map_dfr()

The additional _dfr() tells purrr to return a data frame (df) by row-binding each element together (r)

Source: Blog by Garrick Aden-Buie

Sources