Functions are one of the fundamental building blocks of the R language. They are small pieces of reusable code that can be treated like any other R object.

Remember from the first lecture (First Look), that functions are tools or procedures. They do stuff to objects.

As you have worked through the previous parts of this course, you have used some functions already. Functions are usually characterized by the name of the function followed by parentheses: function_name().

Functions require inputs or objects to do stuff to. They may also have options that modify the actions that are carried out on the objects. Both the objects/data and the options are called arguments, and there are these two different types. We have already used both types of arguments when we used the mean() and sum() functions previously.

Functions also return something, usually another object. This object could be a single value, as in the mean() function; or it could be a list, as in the output of many statistical tests or models.

Let’s look at a few examples.

The mean() function takes a vector of numbers as input, and returns the average of all of the numbers in the input vector. This vector is the data type of argument. Providing arguments to a function is also sometimes called passing arguments to that function. Arguments you want to pass to a function go inside the function’s parentheses. Try passing the argument c(2, 4, 5) to the mean() function.

mean(c(2, 4, 5))
## [1] 3.666667

The mean() function also has options type of arguments. In this case, we can tell the function what to do with missing values (NA’s); we have already worked with this option. Add a single missing value to the above vector and re-run the mean() function, without changing anything else.

mean(c(2, 4, 5, NA))
## [1] NA

We get an ‘NA’ as the answer. In some cases, this might be what we want. In other cases, we may want to know the mean without including the NA’s. Run the mean() function again, and change the options argument to remove the NA, as you have done in the past.

mean(c(2, 4, 5, NA), na.rm = TRUE)
## [1] 3.666667

OK. So, we remember that functions have two kinds of arguments, and that these arguments can have default settings. Defaults are most frequently used for the options arguments, but can also apply to data arguments too.

The Sys.Date() function returns a string representing today’s date. Type Sys.Date() below and see what happens.

Sys.Date()
## [1] "2017-10-04"

If you look at the help file for Sys.Date() and also the code, you will see that there are essentially no arguments. To check the code for any function, just type the name into the Console, without the parentheses. Try that now for the Sys.Date() function.

Sys.Date
## function ()
## as.Date(as.POSIXlt(Sys.time()))
## <bytecode: 0x41bbde0>
## <environment: namespace:base>

Functions like Sys.Date() return a value based on your computer’s environment (in this case, the system time), while other functions manipulate input data in order to compute a return value.

Functions usually take arguments which are variables that the function operates on. For example, the mean() function takes a vector as an argument, like in the case of mean(c(2,6,8)). The mean() function then adds up all of the numbers in the vector and divides that sum by the length of the vector.

In the following question you will be asked to modify a script that will appear as soon as you move on from this question. When you have finished modifying the script, save your changes to the script and type submit() and the script will be evaluated. There will be some comments in the script that opens up, so be sure to read them!

The last R expression to be evaluated in a function will become the return value of that function. We want this function to take one argument, x, and return x without modifying it. Delete the pound sign so that x is returned without any modification. Make sure to save your script before you type submit().

boring_function <- function(x) {
x
}

Now that you’ve created your first function let’s test it! Type: boring_function(‘My first function!’). If your function works, it should just return the string: ‘My first function!’

boring_function('My first function!')
## [1] "My first function!"

Congratulations on writing your first function. By writing functions, you can gain serious insight into how R works. As John Chambers, the creator of R once said:

To understand computations in R, two slogans are helpful: 1. Everything that exists is an object. 2. Everything that happens is a function call.

If you want to see the source code for any function, just type the function name without any arguments or parentheses. Let’s try this out with the function you just created. Type: boring_function to view its source code.

boring_function
## function(x) {
##   x
## }
## <environment: 0x6dbbfe0>

Time to make a more useful function! We’re going to replicate the functionality of the mean() function by creating a function called: my_mean(). Remember that to calculate the average of all of the numbers in a vector you find the sum of all the numbers in the vector, and then divide that sum by the number of numbers in the vector.

Make sure to save your script before you type submit().

# You're free to implement the function my_mean however you want, as long as it
# returns the average of all of the numbers in my_vector.
#
# Hint #1: sum() returns the sum of a vector.
#   Ex: sum(c(1, 2, 3)) evaluates to 6
#
# Hint #2: length() returns the size of a vector.
#   Ex: length(c(1, 2, 3)) evaluates to 3
#
# Hint #3: The mean of all the numbers in a vector is equal to the sum of all of
#          the numbers in the vector divided by the size of the vector.
#
# Note for those of you feeling super clever: Please do not use the mean()
# function while writing this function. We're trying to teach you something
# here!
#
# Be sure to save this script and type submit() in the console after you make

my_mean <- function(my_vector) {
# Remember: the last expression evaluated will be returned!
sum(my_vector)/length(my_vector)
}

Now test out your my_mean() function by finding the mean of the vector c(4, 5, 10).

my_mean(c(4, 5, 10))
## [1] 6.333333

Next, let’s try writing a function with default arguments. You can set default values for a function’s arguments, and this can be useful if you think someone who uses your function will set a certain argument to the same value most of the time.

Make sure to save your script before you type submit().

# Let me show you an example of a function I'm going to make up called
# increment(). Most of the time I want to use this function to increase the
# value of a number by one. This function will take two arguments: "number" and
# "by" where "number" is the digit I want to increment and "by" is the amount I
# want to increment "number" by. I've written the function below.
#
# increment <- function(number, by = 1){
#     number + by
# }
#
# If you take a look in between the parentheses you can see that I've set
# "by" equal to 1. This means that the "by" argument will have the default
# value of 1.
#
# I can now use the increment function without providing a value for "by":
# increment(5) will evaluate to 6.
#
# However if I want to provide a value for the "by" argument I still can! The
# expression: increment(5, 2) will evaluate to 7.
#
# You're going to write a function called "remainder." remainder() will take
# two arguments: "num" and "divisor" where "num" is divided by "divisor" and
# the remainder is returned. Imagine that you usually want to know the remainder
# when you divide by 2, so set the default value of "divisor" to 2. Please be
# sure that "num" is the first argument and "divisor" is the second argument.
#
# Hint #1: You can use the modulus operator %% to find the remainder.
#   Ex: 7 %% 4 evaluates to 3.
#
# Remember to set appropriate default values! Be sure to save this
# script and type submit() in the console after you write the function.

remainder <- function(num, divisor = 2) {
# Remember: the last expression evaluated will be returned!
num %% divisor
}

Let’s do some testing of the remainder function. Run remainder(5) and see what happens.

remainder(5)
## [1] 1

Let’s take a moment to examine what just happened. You provided one argument to the function, and R matched that argument to ‘num’ since ‘num’ is the first argument. The default value for ‘divisor’ is 2, so the function used the default value you provided.

Now let’s test the remainder function by providing two arguments. Type: remainder(11, 5) and let’s see what happens.

remainder(11, 5)
## [1] 1

Once again, the arguments have been matched appropriately.

You can also explicitly specify arguments in a function. When you explicitly designate argument values by name, the ordering of the arguments becomes unimportant. You can try this out by typing: remainder(divisor = 11, num = 5).

remainder(divisor = 11, num = 5)
## [1] 5

As you can see, there is a significant difference between remainder(11, 5) and remainder(divisor = 11, num = 5)!

R can also partially match arguments. Try typing remainder(4, div = 2) to see this feature in action.

remainder(4, div = 2)
## [1] 0

A word of warning: in general you want to make your code as easy to understand as possible. Switching around the orders of arguments by specifying their names or only using partial argument names can be confusing, so use these features with caution!

With all of this talk about arguments, you may be wondering if there is a way you can see a function’s arguments (besides looking at the documentation). Thankfully, you can use the args() function! Type: args(remainder) to examine the arguments for the remainder function.

args(remainder)
## function (num, divisor = 2)
## NULL

You may not realize it but I just tricked you into doing something pretty interesting! args() is a function, remainder() is a function, yet remainder was an argument for args(). Yes it’s true: you can pass functions as arguments! This is a very powerful concept. Let’s write a script to see how it works.

Make sure to save your script before you type submit().

# You can pass functions as arguments to other functions just like you can pass
# data to functions. Let's say you define the following functions:
#
#    num1 + num2
# }
#
# multiply_two_numbers <- function(num1, num2){
#   num1 * num2
# }
#
# some_function <- function(func){
#    func(2, 4)
# }
#
# As you can see we use the argument name "func" like a function inside of
# "some_function()." By passing functions as arguments
# some_function(add_two_numbers) will evaluate to 6, while
# some_function(multiply_two_numbers) will evaluate to 8.
#
# Finish the function definition below so that if a function is passed into the
# "func" argument and some data (like a vector) is passed into the dat argument
# the evaluate() function will return the result of dat being passed as an
# argument to func.
#
# Hints: This exercise is a little tricky so I'll provide a few example of how
# evaluate() should act:
#    1. evaluate(sum, c(2, 4, 6)) should evaluate to 12
#    2. evaluate(median, c(7, 40, 9)) should evaluate to 9
#    3. evaluate(floor, 11.1) should evaluate to 11

evaluate <- function(func, dat){
# Remember: the last expression evaluated will be returned!
func(dat)
}

Let’s take your new evaluate() function for a spin! Use evaluate to find the standard deviation of the vector c(1.4, 3.6, 7.9, 8.8).

evaluate(sd, c(1.4, 3.6, 7.9, 8.8))
## [1] 3.514138

The idea of passing functions as arguments to other functions is an important and fundamental concept in programming.

You may be surprised to learn that you can pass a function as an argument without first defining the passed function. Functions that are not named are appropriately known as anonymous functions.

Let’s use the evaluate function to explore how anonymous functions work. For the first argument of the evaluate function we’re going to write a tiny function that fits on one line. In the second argument we’ll pass some data to the tiny anonymous function in the first argument.

Type the following command and then we’ll discuss how it works: evaluate(function(x){x+1}, 6)

evaluate(function(x){x+1}, 6)
## [1] 7

The first argument is a tiny anonymous function that takes one argument x and returns x+1. We passed the number 6 into this function so the entire expression evaluates to 7.

Try using evaluate() along with an anonymous function to return the first element of the vector c(8, 4, 0). Your anonymous function should only take one argument which should be a variable x.

evaluate(function(x){x[1]}, c(8, 4, 0))
## [1] 8

Now try using evaluate() along with an anonymous function to return the last element of the vector c(8, 4, 0). Your anonymous function should only take one argument which should be a variable x.

evaluate(function(x){x[length(x)]}, c(8, 4, 0))
## [1] 0

For the rest of the course we’re going to use the paste() function frequently. Type ?paste so we can take a look at the documentation for the paste function.

?paste

As you can see the first argument of paste() is ... which is referred to as an ellipsis or simply dot-dot-dot. The ellipsis allows an indefinite number of arguments to be passed into a function. In the case of paste() any number of strings can be passed as arguments and paste() will return all of the strings combined into one string.

Just to see how paste() works, type paste(“Programming”, “is”, “fun!”)

paste("Programming", "is", "fun!")
## [1] "Programming is fun!"

Time to write our own modified version of paste().

Make sure to save your script before you type submit().

# The ellipses can be used to pass on arguments to other functions that are
# used within the function you're writing. Usually a function that has the
# ellipses as an argument has the ellipses as the last argument. The usage of
# such a function would look like:
#
# ellipses_func(arg1, arg2 = TRUE, ...)
#
# In the above example arg1 has no default value, so a value must be provided
# for arg1. arg2 has a default value, and other arguments can come after arg2
# depending on how they're defined in the ellipses_func() documentation.
# Interestingly the usage for the paste function is as follows:
#
# paste (..., sep = " ", collapse = NULL)
#
# Notice that the ellipses is the first argument, and all other arguments after
# the ellipses have default values. This is a strict rule in R programming: all
# arguments after an ellipses must have default values. Take a look at the
# simon_says function below:
#
# simon_says <- function(...){
#   paste("Simon says:", ...)
# }
#
# The simon_says function works just like the paste function, except the
# begining of every string is prepended by the string "Simon says:"
#
# Telegrams used to be peppered with the words START and STOP in order to
# demarcate the beginning and end of sentences. Write a function below called
# telegram that formats sentences for telegrams.
# For example the expression telegram("Good", "morning") should evaluate to:
# "START Good morning STOP"

telegram <- function(...){
paste("START", ..., "STOP")
}

Now let’s test out your telegram function. Use your new telegram function passing in whatever arguments you wish!

telegram()
## [1] "START STOP"

Make sure to save your script before you type submit().

# Let's explore how to "unpack" arguments from an ellipses when you use the
# ellipses as an argument in a function. Below I have an example function that
# is supposed to add two explicitly named arguments called alpha and beta.
#
#   # First we must capture the ellipsis inside of a list
#   # and then assign the list to a variable. Let's name this
#   # variable args.
#
#   args <- list(...)
#
#   # We're now going to assume that there are two named arguments within args
#   # with the names alpha and beta. We can extract named arguments from
#   # the args list by used the name of the argument and double brackets. The
#   # args variable is just a regular list after all!
#
#   alpha <- args[["alpha"]]
#   beta  <- args[["beta"]]
#
#   # Then we return the sum of alpha and beta.
#
#   alpha + beta
# }
#
# Have you ever played Mad Libs before? The function below will construct a
# sentence from parts of speech that you provide as arguments. We'll write most
# of the function, but you'll need to unpack the appropriate arguments from the
# ellipses.

# Do your argument unpacking here!
args <- list(...)
place <- args[["place"]]
noun <- args[["noun"]]

# Don't modify any code below this comment.
# Notice the variables you'll need to create in order for the code below to
# be functional!
paste("News from", place, "today where", adjective, "students took to the streets in protest of the new", noun, "being installed on campus.")
}

Time to use your mad_libs function. Make sure to name the place, adjective, and noun arguments in order for your function to work.

mad_libs(adjective = "", place = "", noun = "")
## [1] "News from  today where  students took to the streets in protest of the new  being installed on campus."

We’re coming to the end of this lesson, but there’s still one more idea you should be made aware of.

You’re familiar with adding, subtracting, multiplying, and dividing numbers in R. To do this you use the +, -, *, and / symbols. These symbols are called binary operators because they take two inputs, an input from the left and an input from the right.

In R you can define your own binary operators. In the next script I’ll show you how.

Make sure to save your script before you type submit().

# The syntax for creating new binary operators in R is unlike anything else in
# R, but it allows you to define a new syntax for your function. I would only
# recommend making your own binary operator if you plan on using it often!
#
# User-defined binary operators have the following syntax:
#      %[whatever]%
# where [whatever] represents any valid variable name.
#
# Let's say I wanted to define a binary operator that multiplied two numbers and
# then added one to the product. An implementation of that operator is below:
#
# "%mult_add_one%" <- function(left, right){ # Notice the quotation marks!
#   left * right + 1
# }
#
# I could then use this binary operator like 4 %mult_add_one% 5 which would
# evaluate to 21.
#
# Write your own binary operator below from absolute scratch! Your binary
# operator must be called %p% so that the expression:
#
#       "Good" %p% "job!"
#
# will evaluate to: "Good job!"

"%p%" <- function(left, right){ # Remember to add arguments!
paste(left, right)
}

You made your own binary operator! Let’s test it out. Paste together the strings: ‘I’, ‘love’, ‘R!’ using your new binary operator.

'I' %p% 'love' %p% 'R!'
## [1] "I love R!"

We’ve come to the end of our lesson! Go out there and write some great functions!

Please submit the log of this lesson to Google Forms so that Simon may evaluate your progress.

1. Your wish is my command