Lesson 6 Functions

In previous lessons, we’ve used various functions (like cat(), print()) without delving deeply into their mechanics. This lesson aims to demystify functions by guiding you through the creation and understanding of custom functions. We’ll also explore the concepts of global and local variable scopes.

6.1 Defining Functions

Defining a function in R follows a syntax similar to variable assignments or conditional statements. We start with an identifier (e.g., add_one) on the left of an assignment operator (<-).

The body of the function follows, beginning with the keyword function, the parameter(s) enclosed in parentheses (e.g., input_value), and the function’s body within curly braces. The value of the last executed statement is automatically returned:

add_one <- function (input_value) {
  output_value <- input_value + 1
  return(output_value) # Explicit return
}

add_one(3)

## [1] 4

Call this function by specifying its identifier and the necessary parameter(s). For instance, in our example add_one(3) returns 4.

6.1.1 Understanding the return() Function

In R, functions automatically return the value of the last expression evaluated. The decision to use the return() function explicitly or rely on implicit return is often guided by the specific context of your work, especially in fields like GIS.

Implicit Return:

In R, if return() is not explicitly used, the function returns the result of the last line of code executed. This is often sufficient, especially in simpler functions. For example:

add_one <- function (input_value) {
  input_value + 1
}
add_one(3)  # Returns 4

## [1] 4

Explicit Return:

Using return(), you can explicitly specify what the function should return. This approach is particularly useful when the function has complex logic, or when you need to return a value before reaching the end of the function. For example:

add_one <- function (input_value) {
  result <- input_value + 1
  return(result)
}
add_one(3)  # Also returns 4

## [1] 4

In both examples, calling add_one(3) returns 4. The explicit use of return() in the second example provides clarity about the intended output of the function.

The choice between implicit and explicit return in R functions can depend on various factors, including the complexity of the function and the coding standards of your team or organization. In GIS and similar fields, it’s important to understand both approaches and adapt to the coding practices of your workplace. Always consider the context and ask, “What is the standard here?”

Early Return Example:

Sometimes, it’s necessary to exit a function early, for instance in error handling:

safe_divide <- function (numerator, denominator) {
  if (denominator == 0) {
    return("Error: Division by zero")
  }
  numerator / denominator
}
safe_divide(10, 0)  # Returns "Error: Division by zero"

## [1] "Error: Division by zero"

So, in GIS and data analysis, adapting to the specific requirements of your project and following established coding standards are key to effective and collaborative work.

6.2 More Parameters

Functions can have multiple parameters, separated by commas.

A function expects as many input values as there are parameters specified in its definition; otherwise, an error is triggered.

The area_rectangle function includes two parameters (height and width), calculates the area by multiplying the inputs, and returns the area as a numeric value:

area_rectangle <- function (height, width) {
  return(height * width)
}
area_rectangle(3, 2)

## [1] 6

Default parameters can enhance a function’s flexibility. For instance, you can modify the area_rectangle function parameters as (height, width = 3). Now, with only one input, the function should return YOUR INPUT * 3. Specifying both parameters overrides the default width. Here’s how it looks:

area_rectangle <- function (height, width = 3) {
  return(height * width)
}

# Calling with one parameter uses the default width of 3
area_rectangle(3)  # Returns 9

## [1] 9

# Calling with two parameters uses the specified width
area_rectangle(3, 2)  # Returns 6

## [1] 6

This approach allows the function to be more flexible and adaptable to different use cases.

6.3 Returning Multiple Values

To return multiple values from a function, encapsulate them in a list. For example, rectangle_metrics calculates and returns both the area and perimeter of a rectangle:

rectangle_metrics <- function (height, width) {
  area <- height * width
  perimeter <- 2 * (height + width)
  
  return(list(area = area, perimeter = perimeter))
}

Retrieve these values using list indices [[1]] and [[2]], or by simply using their names:

metrics <- rectangle_metrics(3, 2)

# Using list indices
cat("Area (list index): ", metrics[[1]], "\n")

## Area (list index):  6

cat("Perimeter (list index): ", metrics[[2]], "\n")

## Perimeter (list index):  10

# Using names
cat("Area (name): ", metrics$area, "\n")

## Area (name):  6

cat("Perimeter (name): ", metrics$perimeter, "\n")

## Perimeter (name):  10

Upon defining a function in R, it appears in the Environment Window of RStudio, similar to a variable. When invoked, the R interpreter executes the function stored in memory.

6.4 Functions and Control Structures

Functions can contain loops and conditional statements. Here’s a function that uses a loop to calculate the factorial of a number (e.g., factorial of 3 = 1 * 2 * 3 = 6):

factorial <- function (input_value) {
  result <- 1
  for (i in 1:input_value) {
    cat("Current:", result, " | i:", i, "\n")
    result <- result * i
  }
  return(result)
}
factorial(3)

## Current: 1  | i: 1 
## Current: 1  | i: 2 
## Current: 2  | i: 3

## [1] 6

The function takes a single numeric value as input, defines a variable named result that is equal to ‘1’ and then creates a loop over all the numbers from 1 (variable result) to the input_value. In the loop, the current value of result is multiplied by the value of the iterator i.

Though technically possible, defining a function within loops or conditional statements is generally avoided.

6.5 Scope

Function parameters are internal variables acting as a bridge between the function and its environment. They are ‘visible’ only within the function’s scope.

The scope of a variable refers to the code region where the variable is accessible.

Understanding scope is crucial for several reasons:

• Avoiding Variable Conflicts: Variables with the same name can exist in different scopes without conflicting with each other. This prevents unexpected behavior caused by variable name overlaps.
• Predictable Behavior: Knowing whether a variable is global or local helps predict its behavior and influence within your script. It clarifies where and how variables can be accessed or modified.
• Resource Management: Local variables are often cleared from memory once the function execution is complete, aiding in efficient resource management in larger scripts or applications.

In R, the scope of variables is defined as follows:

• Global Variables: Defined outside of any function and accessible anywhere in the script, including within functions.
• Local Variables: Defined within a function and accessible only in that function’s scope. This includes function parameters, which act as internal variables bridging the function and its environment.

In the following example, x_value is a global variable, while new_value and input_value are local to the times_x function. Accessing new_value or input_value outside times_x would result in an error, but x_value can be referred to within the function:

x_value <- 10
times_x <- function (input_value) {
  new_value <- input_value * x_value
  return(new_value)
}
times_x(2)

## [1] 20

Understanding these distinctions is essential for writing robust and error-free R code, especially in more complex data analysis or GIS projects.

Referring to external global variables in a function is possible but can be dangerous. At the time of execution, one cannot be sure what the value of the global variable is. For instance, other processes might have changed its value, which affects the behavior of the function. In order to fix this problem, define the variable x_value as a default attribute of function times_x.

See solution!

times_x <- function (input_value, x_value = 10) { return(input_value * x_value) }

These lessons have introduced fundamental R programming concepts. The Base R Cheatsheet offers a concise summary of key operations.