Fortran中的零索引数组是什么？请附示例详解

Hey there! Great question—since Fortran is traditionally 1-indexed for arrays, it’s easy to overlook how to set up zero-indexed ones, especially if you’re coming from languages like C or Python where zero-indexing is the norm. Let me break this down clearly for you.

Unlike some languages that force a fixed indexing scheme, Fortran lets you explicitly define the upper and lower bounds of your arrays. To create a zero-indexed array, you just set the lower bound of the array to 0 when declaring it.

1. 静态零索引数组（编译时确定大小） #

The simplest way is to specify the index range directly in the array declaration using the (lower:upper) syntax.

一维零索引数组示例 #

program zero_index_demo
    implicit none
    ! 声明一个索引从0到4的整数数组（共5个元素）
    integer :: my_arr(0:4)
    integer :: idx

    ! 给数组元素赋值
    do idx = 0, 4
        my_arr(idx) = idx * 10  ! arr(0)=0, arr(1)=10, ..., arr(4)=40
    end do

    ! 打印结果
    print *, "One-dimensional zero-indexed array:"
    do idx = 0, 4
        print *, "my_arr(", idx, ") = ", my_arr(idx)
    end do
end program zero_index_demo

When you run this, it will output each element using its zero-based index—no tricks required.

多维零索引数组示例 #

You can apply the same logic to multi-dimensional arrays, setting the lower bound of any (or all) dimensions to 0:

program multi_dim_zero_index
    implicit none
    ! 二维数组：第一维索引0-2，第二维索引0-1（共3×2=6个元素）
    real :: my_matrix(0:2, 0:1)
    integer :: i, j

    ! 填充数组
    do i = 0, 2
        do j = 0, 1
            my_matrix(i,j) = i + j*0.5
        end do
    end do

    ! 打印矩阵
    print *, "Two-dimensional zero-indexed matrix:"
    do i = 0, 2
        print *, "Row ", i, ": ", my_matrix(i,0), " | ", my_matrix(i,1)
    end do
end program multi_dim_zero_index

2. 动态零索引数组（运行时确定大小） #

If you need an array whose size is determined at runtime, use allocate() and specify the zero-based bounds when allocating:

program dynamic_zero_index
    implicit none
    integer, allocatable :: dyn_arr(:)
    integer :: arr_size, idx

    arr_size = 4  ! 我们想要4个元素，索引0-3
    allocate(dyn_arr(0:arr_size-1))  ! 下界0，上界arr_size-1

    ! 赋值
    do idx = 0, arr_size-1
        dyn_arr(idx) = idx**2  ! arr(0)=0, arr(1)=1, arr(2)=4, arr(3)=9
    end do

    print *, "Dynamic zero-indexed array:"
    print *, dyn_arr

    ! 记得释放动态数组
    deallocate(dyn_arr)
end program dynamic_zero_index

关键注意事项 #

• Zero-indexed arrays are fully compliant with Fortran standards—all array operations (slicing, passing to subroutines, etc.) work exactly as you’d expect, as long as you use the correct index range.
• When passing a zero-indexed array to a subroutine, use assumed-shape arrays (like integer :: arr(:)) if you want the subroutine to handle any index bounds. You can use lbound() and ubound() to get the actual lower/upper bounds of the array inside the subroutine:

  subroutine print_any_array(arr)
      implicit none
      integer :: arr(:)
      integer :: idx, lower, upper
  
      lower = lbound(arr, 1)
      upper = ubound(arr, 1)
  
      print *, "Array bounds: ", lower, " to ", upper
      do idx = lower, upper
          print *, "arr(", idx, ") = ", arr(idx)
      end do
  end subroutine print_any_array
  

  This subroutine will work seamlessly with both 1-indexed and 0-indexed arrays.

内容的提问来源于stack exchange，提问作者Prashant Shukla