使用什么数据结构?(哈希映射vs. trie vs. ?)

I have a C function that produces about 6 million unique arrays. These arrays always have 17 elements each, and each element is an integer from 0 to 16. I also have a slightly modified version of that function that will also produce about 6 million unique arrays of the same kind. My problem is that the second one produces about 45,000 results less than the first, and I'd like to see what these results are.I have a C function that produces about 6 milli