The last major component, and ‘secret weapon’ or ‘unofficial guide’ (according to me) is the the test suite in the official repo (https://github.com/diesel-rs/diesel/tree/master/diesel_tests).

The tests are indeed informative, and give a good idea of how it would be used in practice. I just looked at a few and I have a much better idea of how Diesel looks to be used in practice.

What I didn't see, and don't know if it exists since I haven't read a lot of the docs, is whether Diesel has a built in way to deliver joined results as HashMaps with all fields, as that would be the simplest way I would use it in practice.

For example, if I'm doing the equivalent of:

    SELECT user.*, post.*, comment.* FROM user LEFT JOIN post ON user.id = post.user_id LEFT JOIN comment ON post.id = comment.post_id

Does it have a helper method to identify the three tables that are joined, and instead of returning individual rows, return a HashMap of the user_id to user, a HashMap of the user_id to a list of posts, and a HashMap of the post_id to a list of comments? I think that would be:

    vec![ HashMap<i32,User>, HashMap<i32,Vec<Post>>, HashMap<i32,Vec<Comment>> ]

This would approximate the actual "objectiness" or many ORMs, and would make it easier to loop over data in the desired way with a minimum of boilerplate.

Edit: Or perhaps the more performant way would be to allow registering of callbacks for each change of user, post and comment, and call the appropriate one if defined. That way there isn't time and memory spent on condensing the data into hashmaps? I'm new to thinking about how to deal with ORM results in a language like this.