Whats the difference of functions and methods in Go?

Question

I am trying to get started with Go and the documentation is very good. What I did not find in the documentation is the difference between functions and methods.

As f

Answer 1

As far as I understand at the moment: functions are "global", which means I do not have to import a package to use functions, they are always there. Methods are bound to packages. Is this correct?

No, that's not correct. There are just a couple of functions from the builtin package which are always available. Everything else needs to be imported.

The term "method" came up with object-oriented programming. In an OOP language (like C++ for example) you can define a "class" which encapsulates data and functions which belong together. Those functions inside a class are called "methods" and you need an instance of that class to call such a method.

In Go, the terminology is basically the same, although Go isn't an OOP language in the classical meaning. In Go, a function which takes a receiver is usually called a method (probably just because people are still used to the terminology of OOP).

So, for example:

func MyFunction(a, b int) int {
  return a + b
}
// Usage:
// MyFunction(1, 2)

but

type MyInteger int
func (a MyInteger) MyMethod(b int) int {
  return a + b
}
// Usage:
// var x MyInteger = 1
// x.MyMethod(2)

Answer 2

Tux's answer is great, but I want to augment it with the usage of Go's methods with structs (because this is where I used it often). So let's assume you want to build something to calculate various methods on triangles. You start with a struct:

type Triangle struct {
    a, b, c float64
}

and then you would like to add some functions to calculate the perimeter and square:

func valid(t *Triangle) error {
    if t.a + t.b > t.c && t.a + t.c > t.b && t.b + t.c > t.a {
        return nil
    }
    return errors.New("Triangle is not valid")
}

func perimeter(t *Triangle) (float64, error) {
    err := valid(t)
    if err != nil {
        return -1, err
    }

    return t.a + t.b + t.c, nil
}

func square(t *Triangle) (float64, error) {
    p, err := perimeter(t)
    if err != nil {
        return -1, err
    }

    p /= 2
    s := p * (p - t.a) * (p - t.b) * (p - t.c)
    return math.Sqrt(s), nil
}

And now you got your working program Go Playground. In this case your function takes a parameter (pointer to a triangle) and does something. In OOP word people might have created a class and then added methods. We can see our struct as kind of class with fields and now we add methods:

func (t *Triangle) valid() error {
    if t.a + t.b > t.c && t.a + t.c > t.b && t.b + t.c > t.a {
        return nil
    }
    return errors.New("Triangle is not valid")
}

func (t *Triangle) perimeter() (float64, error) {
    err := t.valid()
    if err != nil {
        return -1, err
    }

    return t.a + t.b + t.c, nil
}

func (t *Triangle) square() (float64, error) {
    p, err := t.perimeter()
    if err != nil {
        return -1, err
    }

    p /= 2
    s := p * (p - t.a) * (p - t.b) * (p - t.c)
    return math.Sqrt(s), nil
}

and we have a fully working example.

Notice that it looks really like a method for objects.

Answer 3

They are explained in detail here - https://anil.cloud/2017/01/26/golang-functions-methods-simplified/

A function in Go follows the syntax:

func FunctionName(Parameters...) ReturnTypes...

Example:

func add(x int, y int) int

To execute:

  add(2,3) 

A method is like a function, but attached to a type (called as receiver). The official guide states “A method is a function with a special receiver argument”. The receiver appears in between the func keyword and the method name. The syntax of a method is:

func (t ReceiverType) FunctionName(Parameters...) ReturnTypes...

Example:

func (t MyType) add(int x, int y) int

To execute:

type MyType string
t1 := MyType("sample")
t1.add(1,2)

Now lets bring pointers into the table. Go lang is pass by value, means fresh copies of the parameters are passed to each function/method call. To pass them by reference you can use pointers.

Syntax of function with pointer in argument/parameter list.

func FunctionName(*Pointers...,Parameters...) ReturnTypes...

Example

func add(t *MyType, x int, y int) int

To execute:

type MyType string
t1 := MyType("sample")
add(&t1,4,5)

Similarly for methods, the receiver type can be a pointer. Syntax of method with pointer (as receiver)

func (*Pointer) FunctionName(Parameters...) ReturnTypes...

Example

func (t *MyType) add(x int, y int) int

To execute:

type MyType string
t1 := MyType("sample")
t1.add(2,3)

Note that we can still write t1.add() to execute the method with a pointer receiver(even-though ‘t1’ is not a pointer) and Go will interpret it as (&t1).add(). Similarly a method with value receiver can be called using pointer too, Go will interpret p.add() as as (*p).add() in that case (where ‘p’ is a pointer). This is applicable only for methods and not for functions.

Methods with pointer receiver are very useful to get a “Java” like behavior where the method is actually modifying on the value to which the receiver points and not on a copy of it.