Functions

Atom

Introduction

It is possible to write your own functions with the function() function. The structure for a function would be:

your_name_for_function <- function(argument 1, argument 2...argument n) {
< the function details >
}

For example, to create a function that will double any number that is put into it,

mydouble <- function(x) {
    x^2
}
mydouble(2)

## [1] 4

mydouble(4.5)

## [1] 20.25

Practice

• Create a function that will raise any input to the power of three
• Make a function called combiner which will add two numbers together.

Re-basing function

In order to create a function that will rebalance any series there are a number of necessary steps:

• find a suitable name
• Write the code to carry out the re-basing
• Wrap that code in a function, thinking about the arguments that will be needed.

Let’s call this myrebase.

myrebase <- function(data = x, base = 1, value = 100){
    mydata <- x/x[base] * value
    return(mydata)
}

Notice that I have added return() to the end of the function. This is not necessary as the function will return (or preserve) the last object that was created (mydata in this case). However, it is good practice because it will make you think what you want to be returned. Everything else will disappear after the function run has been completed.

Now we need to test the function to make sure it works. Make some toy data and run it in different ways.

x <- 1:3
myrebase(x)

## [1] 100 200 300

myrebase(x, base = 3)

## [1]  33.33333  66.66667 100.00000

myrebase(x, base = 3, value = 3)

## [1] 1 2 3

myrebase(x, base = 3, value = 1)

## [1] 0.3333333 0.6666667 1.0000000

Percentage change function

This is a function that will calculate the percentage change or return.

mypercent <- function(x, reverse  = FALSE, lag = 1){
    if(reverse == TRUE){
        myreturn <- x[1:(length(x) - lag)] /
            x[(lag + 1):length(x)] - 1
        myvector <- c(myreturn, rep(NA, lag))
    } else {
        myreturn <- x[(lag + 1):length(x)] /
            x[1:(length(x) - lag)] - 1
        myvector <- c(rep(NA, lag), myreturn)
    }
    return(myvector)
}

There is a lot here but it is a very versatile function.

• it can deal with data that runs from top to bottom or the reverse: bottom to top
• it makes use of the if() function. If we tell the function that the data is in reverse it will adjust the calculation.
• it will pad the returned vector of returns so that it can be immediately connected to an existing data frame. We also need to test the function to make sure that it works.

x <- 1:3
mypercent(x)

## [1]  NA 1.0 0.5

mypercent(x, reverse = TRUE)

## [1] -0.5000000 -0.3333333         NA

mypercent(x, lag = 2)

## [1] NA NA  2

mypercent(x, reverse = TRUE, lag = 2)

## [1] -0.6666667         NA         NA

Practice

• See if you can add the option to choose a particular date for the re-base function

Making functions available

Once you have functions you can put them in a single folder and call it at the start of any work by using the

source('./R/functions.R')

Note, './R/functions.R' must reflect the place where the functions are compared to where you are now.