问题描述
我知道对于 associative 数组，数组键没有固有的顺序:
I know that for associative arrays, there's no inherent order for the array keys:
declare -A map=([a]=b [c]=d [e]=f)
echo "${!map[@]}" # => e c a (perhaps)
那索引数组呢?
尝试一下:
declare -a list
for i in {1..1000}; do list[RANDOM]=1; done
echo "${!list[@]}"
看来索引是按数字顺序排列的.但是我可以依靠吗?
It appears that the indices are numerically ordered. But can I rely on that?
推荐答案
我花了一些时间阅读 bash来源.索引数组被实现为双链表.
I spent some time reading the bash source. Indexed arrays are implemented as doubly-linked lists.
插入新的数组元素( ary [i] = value )将遍历列表并插入元素，以使索引保持数字排序: https://git.savannah.gnu.org/cgit/bash.git/tree/array.c#n548
Inserting new array elements (ary[i]=value) will walk the list and insert the element such that the indices remain numerically sorted: https://git.savannah.gnu.org/cgit/bash.git/tree/array.c#n548
提取索引( $ {！ary [@]} )将列表从头(具有最大索引的元素)移至尾部(最小索引)，并建立一个列表索引: https://git.savannah.gnu.org/cgit/bash.git/tree/array.c#n778
Extracting the indices (${!ary[@]}) walks the list from the head (the element with the max index) back to the tail (min index) and build a list of indices: https://git.savannah.gnu.org/cgit/bash.git/tree/array.c#n778
因此，目前已实施，我想我可以确信索引列表的确是有序的.
So, as currently implemented, I think I can feel confident that the list of indices is indeed ordered.
这篇关于从索引数组中提取索引:保证要排序?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！