深入理解Go语言中的数组和切片

一、类型

数组是值类型，将一个数组赋值给另一个数组时，传递的是一份拷贝。

切片是引用类型，切片包装的数组称为该切片的底层数组。

我们来看一段代码

//a是一个数组，注意数组是一个固定长度的，初始化时候必须要指定长度，不指定长度的话就是切片了
a := [3]int{1, 2, 3}
//b是数组，是a的一份拷贝
b := a
//c是切片，是引用类型，底层数组是a
c := a[:]
for i := 0; i < len(a); i++ {
 a[i] = a[i] + 1
}
//改变a的值后，b是a的拷贝，b不变，c是引用，c的值改变
fmt.println(a) //[2,3,4]
fmt.println(b) //[1 2 3]
fmt.println(c) //[2,3,4]

二、make

make 只能用于slice, map 和 channel, 所以下面一段代码生成了一个slice，是引用类型

s1 := make([]int, 0, 3)

for i := 0; i < cap(s1); i++ {
 s1 = append(s1, i)
}
s2 := s1
for i := 0; i < len(a); i++ {
 s1[i] = s1[i] + 1
}

fmt.println(s1) //[1 2 3]
fmt.println(s2) //[1 2 3]

三、当对slice append 超出底层数组的界限时

//n1是n2的底层数组
n1 := [3]int{1, 2, 3}
n2 := n1[0:3]
fmt.println("address of items in n1: ")
for i := 0; i < len(n1); i++ {
 fmt.printf("%p\n", &n1[i])
}
//address of items in n1:
//0xc20801e160
//0xc20801e168
//0xc20801e170
fmt.println("address of items in n2: ")
for i := 0; i < len(n2); i++ {
 fmt.printf("%p\n", &n2[i])
}
//address of items in n2:
//0xc20801e160
//0xc20801e168
//0xc20801e170

//对n2执行append操作后，n2超出了底层数组n1的j
n2 = append(n2, 1)
fmt.println("address of items in n1: ")
for i := 0; i < len(n1); i++ {
 fmt.printf("%p\n", &n1[i])
}
//address of items in n1:
//0xc20801e160
//0xc20801e168
//0xc20801e170

fmt.println("address of items in n2: ")
for i := 0; i < len(n2); i++ {
 fmt.printf("%p\n", &n2[i])
}
//address of items in n2:
//0xc20803a2d0
//0xc20803a2d8
//0xc20803a2e0
//0xc20803a2e8

四、引用“失效”

实现了删除slice最后一个item的函数

func rmlast(a []int) {
 fmt.printf("[rmlast] the address of a is %p", a)
 a = a[:len(a)-1]
 fmt.printf("[rmlast] after remove, the address of a is %p", a)
}

调用此函数后，发现原来的slice并没有改变

func main() {
 xyz := []int{1, 2, 3, 4, 5, 6, 7, 8, 9}
 fmt.printf("[main] the address of xyz is %p\n", xyz)
 rmlast(xyz)
 fmt.printf("[main] after remove, the address of xyz is %p\n", xyz)
 fmt.printf("%v", xyz) //[1 2 3 4 5 6 7 8 9]
}

打印出来的结果如下：

[main] the address of xyz is 0xc2080365f0
[rmlast] the address of a is 0xc2080365f0
[rmlast] after remove, the address of a is 0xc2080365f0
[main] after remove, the address of xyz is 0xc2080365f0
[1 2 3 4 5 6 7 8 9]

这里直接打印了slice的指针值，因为slice是引用类型，所以指针值都是相同的，我们换成打印slice的地址看下

func rmlast(a []int) {
 fmt.printf("[rmlast] the address of a is %p", &a)
 a = a[:len(a)-1]
 fmt.printf("[rmlast] after remove, the address of a is %p", &a)
}
func main() {
 xyz := []int{1, 2, 3, 4, 5, 6, 7, 8, 9}
 fmt.printf("[main] the address of xyz is %p\n", &xyz)
 rmlast(xyz)
 fmt.printf("[main] after remove, the address of xyz is %p\n", &xyz)
 fmt.printf("%v", xyz) //[1 2 3 4 5 6 7 8 9]
}

结果：

[main] the address of xyz is 0xc20801e1e0
[rmlast] the address of a is 0xc20801e200
[rmlast] after remove, the address of a is 0xc20801e200
[main] after remove, the address of xyz is 0xc20801e1e0
[1 2 3 4 5 6 7 8 9]

这次可以看到slice作为函数参数传入函数时，实际上也是拷贝了一份slice，因为slice本身是个指针，所以从现象来看，slice是引用类型

总结

以上就是这篇文章的全部内容，希望对大家的学习或者工作带来一定的帮助，如果有疑问大家可以留言交流。