Otherwise there are constructions that provide a virtual copy of an iterator and cache the values they still need for the second loop.

Or just don't use an iterator if it doesn't fit the model you need, or turn it into an array before.

(Part of) the idea behind generators is that you don't store all results, because they could be a long or possibly even infinite sequence. E.g. in Python the basic iterator is range(X), which returns the numbers from 0 to X-1. If X=100000, it is a bad idea to generate a giant list of all those numbers, just because you want to know how often you ran through a loop.