Functions

Learning objectives

• Questions:

  • How to write functions in R

• Objectives:

  • Define functions

  • Return value(s) from functions

• Keypoints:

  • Use function to define a new function in R

  • Use parameters to pass value to function

  • Load function into program using source()

• We have seen some examples of built-in R functions. For some functions, you would have to install particular packages. In this chapter, we will show you how to write your own functions.

Using custom functions

• There are several in-built functions in R that can be used to perform analytical tasks, for example: mean, min, max, quantile,summary.

• For example, here’s the mean function, which computes the arithmetic average of a vector:

• Using function mean with missing value

mean (c (1, 3, 5, 3, 2))
v <- c(2,NA,4,NaN,6)
mean (v)
mean(v,na.rm=TRUE)

Writing a user-defined function with 1 argument

• Syntax:

f <- function(arg){
  do function with argument
}

• Example:

squareroot <- function(a){
  a^0.5
}
squareroot(49)

Writing a user-defined function with 2 or more arguments

• Syntax:

f <- function(arg1,arg2){
  do function with arg1 & arg2
}

• Example:

Addtwo <- function(a,b){
  a+b
}
Addtwo(1,2)

Specifying a variable for the result

• Syntax:

f <- function(args){
  f1 <- do function with args
  return(f1)
}

• For example:

# Function to convert oF to oC
F2C <- function(temp){
   c <- ((temp - 32) * (5 / 9))
   return(c)
}
F2C(100)

Returning several results in a list

• Syntax:

f <- function(args){
  do function with args
  out1 <- do1
  out2 <- do2  
  output <- list(out1=out1,out2=out2)
}

• Example: a function which converts polar coordinates to Cartesian coordinates

polar2cart <- function (r, phi) {
  x <- r*sin(phi)
  y <- r*cos(phi)
  return (list(x, y))
}
polar2cart(1,pi/6)[1] 
polar2cart(1,pi/6)[2] 

• Let’s specify the names of the two outputs:

polar2cart <- function (r, phi) {
  xcoord <- r*sin(phi)
  ycoord <- r*cos(phi)
  return (list(x=xcoord, y=ycoord))
}
polar2cart(1,pi/6)[1] 
polar2cart(1,pi/6)[2] 
polar2cart(1,pi/6)$x
polar2cart(1,pi/6)$y 

Nested functions

• In complex data science use cases, we may have to work on nested functions, which contain functions within a function.

• For example: Given dataset mtcars. Find the mean of fuel consumption mpg for cars that having 4 cylinders cyl

data(mtcars)
names(mtcars)

# Step 1: find the cars that having 4 cylinders:
ind <- mtcars$cyl==4
# Step 2: find the fuel consumption of all the cars having 4 cylinders:
fuel_4cyl <- mtcars$mpg[ind]
# Step 3: compute the mean
mean(fuel_4cyl)

• All the 3 steps can be nested into one command line for experience user:

mean(mtcars$mpg[mtcars$cyl==4])

Defensive programming with stopifnot() function

• Defensive programming encourages us to frequently check conditions and throw an error if something is wrong.

• For example:

F2C <- function(temp){
   stopifnot(is.numeric(temp)==TRUE)
   c <- ((temp - 32) * (5 / 9))
   return(c)
}
F2C(100a)
F2C(100)

Saving functions for future use

• Let’s save our function so we can use it later.

• First, get your working directory by running getwd(). Then, in a file browser (Windows) or in Finder (Mac), go to that directory, and please create a folder “R_workshop”. Then, in R, let’s make this our working directory: setwd ("R_workshop"). Then, let’s list the files in it:

list.files (getwd())

• The result should be empty (character(0)) because it is an empty directory. If you are familiar with Linux, you can use the Terminal tab from the console for the same purpose.

• Then, in the R studio Editor, copy-and-paste the polar2cart function:

polar2cart <- function (r, phi) {
  x <- r*sin(phi)
  y <- r*cos(phi)
  return (list(x, y))
}

…and then save it by doing File -> Save As and selecting the name polar2cart.R. Make sure you are saving it in the R_workshop folder.