[This is a follow up to a previous question of mine,](https://stackapps.com/questions/9226/how-can-i-retrieve-the-reputation-i-or-another-user-fwiw-earned-for-a-given-t) and the starting point is that, given a list of [reputation objects](https://api.stackexchange.com/docs/types/reputation) retrieved via [`/users/{ids}/reputation`](https://api.stackexchange.com/docs/reputation-on-users), I can easily retain only the entries with `post_type == 'answers'` and then get `post_id` for each of them, ending up with a list of `answer_id`s. Given this list of `answer_id`s, how can I efficiently obtain the list of [`question` objects](https://api.stackexchange.com/docs/types/question) associated to those `answer_id`s **_without_** loosing the information of which `question` comes from which `answer_id`? Clearly, I could make one [`/answers/{ids}/questions`](https://api.stackexchange.com/docs/questions-by-answer-ids) request for each of the `answer_id`s, but this means... I'm being kicked out with a `503 (Service unavailable)` (hence the [tag:limits] ta). On the other hand, making one single `/answers/{ids}/questions` request for all the `ids` (well, I should actually make a request for chunks of 100 entries; then maybe I have to browse the few resulting `page`s; but this probably still means a small number of requests), I'm a bit unsure about whether those `question`s are obtained in the same order as the the `answer_id`s. Quite the opposite, from [this answer](https://stackapps.com/a/6295/85941) I understand that there's no guarantee at all that that's the case. If the retrieved `question`s are not sorted as the `answer_id`s, then the only way I see to re-estabilish the relation is to make a [`/answers/{ids}`](https://api.stackexchange.com/docs/answers-by-ids) request as well (again, in chunks of 100 entries, ...), so that I can recognize which `question` corresponds to which `answer_id`, as the `question_id` is both in `question`s and [`answer`s](https://api.stackexchange.com/docs/types/answer). But this looks a bit a convoluted approach. Plus, retrieving the whole `answer` objects when all I need is the `question_id` for each of them seems overkill. Is there any other strategy I could adopt? Or is the latter one I described (retrieving all `question`s and all `answer`s from the `answer_id`s) just fine?