# Functions

## Overview

Teaching: 30 min
Exercises: 10 min
Questions
• What is a function?

• How do functions help us with reproducibility?

• How do I write a function in `R`?

Objectives

• Learn how to write a function in `R`

## Overview

What is a function?

``````input --> function does something --> output
``````

One of the main benefits of writing functions is to avoid repetition of your code. When you write code you should aim for laziness. If you are about to copy a piece of code 5 times and just change a variable in each instance, you are better off converting it into a function. In general, you should never repeat yourself when you write code, it’s called the DRY (Don’t Repeat Yourself) principle.

Another advantage is that your code is self contained and any variable that you create inside a function will not be exported into your global environment.

## How to write functions in R?

#### Let’s convert Fahrenheit to Celsius

``````(70 - 32) * (5 / 9)
``````
``````##  21.11111
``````
``````(65 - 32) * (5 / 9)
``````
``````##  18.33333
``````
``````(85 - 32) * (5 / 9)
``````
``````##  29.44444
``````

#### Let’s convert this into a `function`:

``````fahr_to_celsius <- function(temp) {
(temp - 32) * (5 / 9)
}
``````

### Don’t forget to re-evaluate your function, after modifying it.

#### A little better:

``````convert_fahr <- function(temp, to) {
res <- (temp - 32) * (5 / 9)
if (to == "kelvin") {
res <- res + 273.15
}
res
}
``````

#### With functions you can easily control the format of the input and avoid the chances for typos or other small mistakes.

``````convert_fahr <- function(temp, to = c("celsius", "kelvin")) {
to <- tolower(to)
to <- match.arg(to)
res <- (temp - 32) * (5 / 9)
if (to == "kelvin") {
res <- res + 273.15
}
res
}
``````

#### Let’s refactor this function into something even better that will allow you to easily expend the convert_fahr function to other units:

``````fahr_to_celsius <- function(temp) {
(temp - 32) * (5 / 9)
}

celsius_to_kelvin <- function(temp) {
temp + 273.15
}

convert_fahr <- function(temp, to = c("celsius", "kelvin")) {
to <- tolower(to)
to <- match.arg(to)
res <- fahr_to_celsius(temp)
if (to == "kelvin") {
res <- celsius_to_kelvin(res)
}
res
}
``````

Your intentions are clear, everything is self contained, you can easily debug, test and document each of the steps involved.

## Converter challenge

Write a function that converts pounds in kilograms (divides pounds by 2.2).

Stretch goal: and in grams

## Solution

This is the body of the solution.

``````lb_to_kg <- function(pounds){
kg <- pounds * 2.2
return(kg)
}
``````

In grams:

``````lb_to_kg <- function(pounds){
gm <- pounds * 2.2 * 1000
return(gm)
}
``````

## Key Points

• Functions complete tasks in a reproducible manner

• Functions improve reproducibility