https://segmentfault.com/a/

通过下面几个问题来更好理解golang 的数组和切片

  1. 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

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    //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]

  2. 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

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    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]

  3. 当对slice append 超出底层数组的界限时

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    28
    29
    30
    31
    32
    33
    34
    35
    36
    37
    38
    39
    40
    //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

  4. 引用“失效”
    实现了删除slice最后一个item的函数

    1
    2
    3
    4
    5
    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

    1
    2
    3
    4
    5
    6
    7
    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：

    1
    2
    3
    4
    5
    [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

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    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：

    1
    2
    3
    4
    5
    [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