在Go中什么时候使用指针？

1 在方法中使用指针

什么是receiver？

func (t T) method_name(t T){}

这里面的T就是receiver

• 使用receiver作为方法参数

func main() {
   r := receiver{Name: "zs"}
   fmt.Println(r)
   r.methodA()
   fmt.Println(r)
}

type receiver struct {
   Id   int
   Name string
   Age  int
}

func (receiver receiver) methodA() {
   receiver.Name = "ls"
}

func (receiver *receiver) methodB() {
   receiver.Name = "ls"
}

结果：

{0 zs 0}
{0 zs 0}

• 使用*receiver作为方法参数

func main() {
   r := receiver{Name: "zs"}
   fmt.Println(r)
   r.methodB()
   fmt.Println(r)
}

type receiver struct {
   Id   int
   Name string
   Age  int
}

func (receiver receiver) methodA() {
   receiver.Name = "ls"
}

func (receiver *receiver) methodB() {
   receiver.Name = "ls"
}

结果：

{0 zs 0}
{0 ls 0}

2 在结构体中使用指针

• 方式一

func main() {
   student := Student{Map: map[string]int{"S": 0}, ReceiverA: receiver{Name: "A"}, ReceiverB: &receiver{Name: "B"}}
   fmt.Println(student, *student.ReceiverB)
   student.updateA()
   fmt.Println(student, *student.ReceiverB)
}

type receiver struct {
   Id   int
   Name string
   Age  int
}

type Student struct {
   No        int
   Map       map[string]int
   ReceiverA receiver
   ReceiverB *receiver
}

func (stu Student) updateA() {
   stu.Map["a"] = 1
   stu.ReceiverA = receiver{Name: "ww"}
   stu.ReceiverB = &receiver{Name: "ww"}
}

func (stu *Student) updateB() {
   stu.Map["b"] = 2
   stu.ReceiverA = receiver{Name: "ww"}
   stu.ReceiverB = &receiver{Name: "ww"}
}

结果：

{0 map[S:0] {0 A 0} 0xc0000b4000} {0 B 0}
{0 map[S:0 a:1] {0 A 0} 0xc0000b4000} {0 B 0}

• 方式二

func main() {
   student := Student{Map: map[string]int{"S": 0}, ReceiverA: receiver{Name: "A"}, ReceiverB: &receiver{Name: "B"}}
   fmt.Println(student, *student.ReceiverB)
   student.updateA()
   fmt.Println(student, *student.ReceiverB)
}

type receiver struct {
   Id   int
   Name string
   Age  int
}

type Student struct {
   No        int
   Map       map[string]int
   ReceiverA receiver
   ReceiverB *receiver
}

func (stu Student) updateA() {
   stu.Map["a"] = 1
   stu.ReceiverA = receiver{Name: "ww"}
   stu.ReceiverB = &receiver{Name: "ww"}
}

func (stu *Student) updateB() {
   stu.Map["b"] = 2
   stu.ReceiverA = receiver{Name: "ww"}
   stu.ReceiverB = &receiver{Name: "ww"}
}

结果：

{0 map[S:0] {0 A 0} 0xc0000b4000} {0 B 0}
{0 map[S:0 b:2] {0 ww 0} 0xc0000b4060} {0 ww 0}

3 什么时候使用指针

一个函数何时该用指针类型做receiver对初学者而言一直是个头疼的问题。下面是列举了一些常用的判断指导。

• 如果receiver是map、func或者chan，不要使用指针
• 如果receiver是slice并且该函数并不会修改此slice，不要使用指针
• 如果该函数会修改receiver，此时一定要用指针
• 如果receiver是struct并且包含互斥类型sync.Mutex，或者是类似的同步变量，receiver必须是指针，这样可以避免对象拷贝
• 如果receiver是较大的struct或者array，使用指针则更加高效。多大才算大？假设struct内所有成员都要作为函数变量传进去，如果觉得这时数据太多，就是struct太大
• 如果receiver是struct，array或者slice，并且其中某个element指向了某个可变量，则这个时候receiver选指针会使代码的意图更加明显
• 如果receiver使较小的struct或者array，并且其变量都是些不变量、常量，例如time.Time，value receiver更加适合，因为value receiver可以减少需要回收的垃圾量。
• 最后，如果不确定用哪个，使用指针类的receiver

参考文章：

https://zhuanlan.zhihu.com/p/395747448