Writing flexible, reusable, and modular code is vital for developing versatile programs. Working in this way ensures code is easier to maintain by avoiding the need to make the same change in multiple places. How you accomplish this varies from language to language. For instance, inheritance is a common approach that is used in languages such as Java, C++, C#, and more.
Developers can also attain those same design goals through composition. Composition is a way to combine objects or data types into more complex ones. This is the approach that Go uses to promote code reuse, modularity, and flexibility. Interfaces in Go provide a method of organizing complex compositions, and learning how to use them will allow you to create common, reusable code.
In this article, we will learn how to compose custom types that have common behaviors, which will allow us to reuse our code. We’ll also learn how to implement interfaces for our own custom types that will satisfy interfaces defined from another package.
One of the core implementations of composition is the use of interfaces. An interface defines a behavior of a type. One of the most commonly used interfaces in the Go standard library is the fmt.Stringer
interface:
type Stringer interface {
String() string
}
The first line of code defines a type
called Stringer
. It then states that it is an interface
. Just like defining a struct, Go uses curly braces ({}
) to surround the definition of the interface. In comparison to defining structs, we only define the interface’s behavior; that is, “what can this type do”.
In the case of the Stringer
interface, the only behavior is the String()
method. The method takes no arguments and returns a string.
Next, let’s look at some code that has the fmt.Stringer
behavior:
package main
import "fmt"
type Article struct {
Title string
Author string
}
func (a Article) String() string {
return fmt.Sprintf("The %q article was written by %s.", a.Title, a.Author)
}
func main() {
a := Article{
Title: "Understanding Interfaces in Go",
Author: "Sammy Shark",
}
fmt.Println(a.String())
}
The first thing we do is create a new type called Article
. This type has a Title
and an Author
field and both are of the string data type:
...
type Article struct {
Title string
Author string
}
...
Next, we define a method
called String
on the Article
type. The String
method will return a string that represents the Article
type:
...
func (a Article) String() string {
return fmt.Sprintf("The %q article was written by %s.", a.Title, a.Author)
}
...
Then, in our main
function, we create an instance of the Article
type and assign it to the variable called a
. We provide the values of "Understanding Interfaces in Go"
for the Title
field, and "Sammy Shark"
for the Author
field:
...
a := Article{
Title: "Understanding Interfaces in Go",
Author: "Sammy Shark",
}
...
Then, we print out the result of the String
method by calling fmt.Println
and passing in the result of the a.String()
method call:
...
fmt.Println(a.String())
After running the program you’ll see the following output:
OutputThe "Understanding Interfaces in Go" article was written by Sammy Shark.
So far, we haven’t used an interface, but we did create a type that had a behavior. That behavior matched the fmt.Stringer
interface. Next, let’s see how we can use that behavior to make our code more reusable.
Now that we have our type defined with the desired behavior, we can look at how to use that behavior.
Before we do that, however, let’s look at what we would need to do if we wanted to call the String
method from the Article
type in a function:
package main
import "fmt"
type Article struct {
Title string
Author string
}
func (a Article) String() string {
return fmt.Sprintf("The %q article was written by %s.", a.Title, a.Author)
}
func main() {
a := Article{
Title: "Understanding Interfaces in Go",
Author: "Sammy Shark",
}
Print(a)
}
func Print(a Article) {
fmt.Println(a.String())
}
In this code we add a new function called Print
that takes an Article
as an argument. Notice that the only thing the Print
function does is call the String
method. Because of this, we could instead define an interface to pass to the function:
package main
import "fmt"
type Article struct {
Title string
Author string
}
func (a Article) String() string {
return fmt.Sprintf("The %q article was written by %s.", a.Title, a.Author)
}
type Stringer interface {
String() string
}
func main() {
a := Article{
Title: "Understanding Interfaces in Go",
Author: "Sammy Shark",
}
Print(a)
}
func Print(s Stringer) {
fmt.Println(s.String())
}
Here we created an interface called Stringer
:
...
type Stringer interface {
String() string
}
...
The Stringer
interface has only one method, called String()
that returns a string
. A method is a special function that is scoped to a specific type in Go. Unlike a function, a method can only be called from the instance of the type it was defined on.
We then update the signature of the Print
method to take a Stringer
, and not a concrete type of Article
. Because the compiler knows that a Stringer
interface defines the String
method, it will only accept types that also have the String
method.
Now we can use the Print
method with anything that satisfies the Stringer
interface. Let’s create another type to demonstrate this:
package main
import "fmt"
type Article struct {
Title string
Author string
}
func (a Article) String() string {
return fmt.Sprintf("The %q article was written by %s.", a.Title, a.Author)
}
type Book struct {
Title string
Author string
Pages int
}
func (b Book) String() string {
return fmt.Sprintf("The %q book was written by %s.", b.Title, b.Author)
}
type Stringer interface {
String() string
}
func main() {
a := Article{
Title: "Understanding Interfaces in Go",
Author: "Sammy Shark",
}
Print(a)
b := Book{
Title: "All About Go",
Author: "Jenny Dolphin",
Pages: 25,
}
Print(b)
}
func Print(s Stringer) {
fmt.Println(s.String())
}
We now add a second type called Book
. It also has the String
method defined. This means it also satisfies the Stringer
interface. Because of this, we can also send it to our Print
function:
OutputThe "Understanding Interfaces in Go" article was written by Sammy Shark.
The "All About Go" book was written by Jenny Dolphin. It has 25 pages.
So far, we have demonstrated how to use just a single interface. However, an interface can have more than one behavior defined. Next, we’ll see how we can make our interfaces more versatile by declaring more methods.
One of the core tenants of writing Go code is to write small, concise types and compose them up to larger, more complex types. The same is true when composing interfaces. To see how we build up an interface, we’ll first start by defining only one interface. We’ll define two shapes, a Circle
and Square
, and they will both define a method called Area
. This method will return the geometric area of their respective shapes:
package main
import (
"fmt"
"math"
)
type Circle struct {
Radius float64
}
func (c Circle) Area() float64 {
return math.Pi * math.Pow(c.Radius, 2)
}
type Square struct {
Width float64
Height float64
}
func (s Square) Area() float64 {
return s.Width * s.Height
}
type Sizer interface {
Area() float64
}
func main() {
c := Circle{Radius: 10}
s := Square{Height: 10, Width: 5}
l := Less(c, s)
fmt.Printf("%+v is the smallest\n", l)
}
func Less(s1, s2 Sizer) Sizer {
if s1.Area() < s2.Area() {
return s1
}
return s2
}
Because each type declares the Area
method, we can create an interface that defines that behavior. We create the following Sizer
interface:
...
type Sizer interface {
Area() float64
}
...
We then define a function called Less
that takes two Sizer
and returns the smallest one:
...
func Less(s1, s2 Sizer) Sizer {
if s1.Area() < s2.Area() {
return s1
}
return s2
}
...
Notice that we not only accept both arguments as the type Sizer
, but we also return the result as a Sizer
as well. This means that we no longer return a Square
or a Circle
, but the interface of Sizer
.
Finally, we print out what had the smallest area:
Output{Width:5 Height:10} is the smallest
Next, let’s add another behavior to each type. This time we’ll add the String()
method that returns a string. This will satisfy the fmt.Stringer
interface:
package main
import (
"fmt"
"math"
)
type Circle struct {
Radius float64
}
func (c Circle) Area() float64 {
return math.Pi * math.Pow(c.Radius, 2)
}
func (c Circle) String() string {
return fmt.Sprintf("Circle {Radius: %.2f}", c.Radius)
}
type Square struct {
Width float64
Height float64
}
func (s Square) Area() float64 {
return s.Width * s.Height
}
func (s Square) String() string {
return fmt.Sprintf("Square {Width: %.2f, Height: %.2f}", s.Width, s.Height)
}
type Sizer interface {
Area() float64
}
type Shaper interface {
Sizer
fmt.Stringer
}
func main() {
c := Circle{Radius: 10}
PrintArea(c)
s := Square{Height: 10, Width: 5}
PrintArea(s)
l := Less(c, s)
fmt.Printf("%v is the smallest\n", l)
}
func Less(s1, s2 Sizer) Sizer {
if s1.Area() < s2.Area() {
return s1
}
return s2
}
func PrintArea(s Shaper) {
fmt.Printf("area of %s is %.2f\n", s.String(), s.Area())
}
Because both the Circle
and the Square
type implement both the Area
and String
methods, we can now create another interface to describe that wider set of behavior. To do this, we’ll create an interface called Shaper
. We’ll compose this of the Sizer
interface and the fmt.Stringer
interface:
...
type Shaper interface {
Sizer
fmt.Stringer
}
...
Note: It is considered idiomatic to try to name your interface by ending in er
, such as fmt.Stringer
, io.Writer
, etc. This is why we named our interface Shaper
, and not Shape
.
Now we can create a function called PrintArea
that takes a Shaper
as an argument. This means that we can call both methods on the passed in value for both the Area
and String
method:
...
func PrintArea(s Shaper) {
fmt.Printf("area of %s is %.2f\n", s.String(), s.Area())
}
If we run the program, we will receive the following output:
Outputarea of Circle {Radius: 10.00} is 314.16
area of Square {Width: 5.00, Height: 10.00} is 50.00
Square {Width: 5.00, Height: 10.00} is the smallest
We have now seen how we can create smaller interfaces and build them up into larger ones as needed. While we could have started with the larger interface and passed it to all of our functions, it is considered best practice to send only the smallest interface to a function that is needed. This typically results in clearer code, as anything that accepts a specific smaller interface only intends to work with that defined behavior.
For example, if we passed Shaper
to the Less
function, we may assume that it is going to call both the Area
and String
methods. However, since we only intend to call the Area
method, it makes the Less
function clear as we know that we can only call the Area
method of any argument passed to it.
We have seen how creating smaller interfaces and building them up to larger ones allows us to share only what we need to a function or method. We also learned that we can compose our interfaces from other interfaces, including those defined from other packages, not just our packages.
If you’d like to learn more about the Go programming language, check out the entire How To Code in Go series.
Thanks for learning with the DigitalOcean Community. Check out our offerings for compute, storage, networking, and managed databases.
Go (or GoLang) is a modern programming language originally developed by Google that uses high-level syntax similar to scripting languages. It is popular for its minimal syntax and innovative handling of concurrency, as well as for the tools it provides for building native binaries on foreign platforms.
This textbox defaults to using Markdown to format your answer.
You can type !ref in this text area to quickly search our full set of tutorials, documentation & marketplace offerings and insert the link!
Great tutorial. In the last large code block for main.go, I think you forgot to highlight the Shaper interface addition.