Hello!

​

Thinking about optimizing my current application. The current schema looks roughly like this.

Two main collections: Project and Album.

Project currently has 200 000 documents, and looks like this, where the reviews array is being updated with new items daily.

{

    "_id" : ObjectId("6064c95822796579f33bea4f"),

    "name" : "justSomeName"

    "reviews" : [

        {

            "album" : ObjectId("5f34ee8bf0857e55ed5ebf97"),

            "notes" : "Crap",

            "rating" : 1

        },

        {

            "album" : ObjectId("5f34ee8bf0857e55ed5ec035"),

            "notes" : "great!",

            "rating" : 5

        },

        {

            .... // Possibbly 1000 more items in this array

        }

    ]

}

​

Album has approximately 1000 documents (and will stay that many), all of them more or less static :

​

{

    "_id" : ObjectId("5f34ee8bf0857e55ed5ebd9c"),

    "name" : "More Songs About Buildings And Food",

    "release_date" : "1978",

    "spotify" : "spotify:album:01RJdKvXyz515O37itqMIJ",

    "genres" : [

        "post-punk",

        "new-wave"

    ]

}

​

The most common query is to get a project with all its reviewed items, so at first i felt this schema structure is a good approach. Keeping the schema tightly connected to how the frontend is consuming it.

Theres no need for getting a single review item.

​

However, now i would need a way of getting all reviews for a specific album, so i guess the best approach here would be to put reviews in its own collection Review, making it easier to query reviews by project and/or by album.

​

Now to the questions:

​

1. Am i right in thinking this is the best solution to my problem? Or is there another, better, way of solving this?
2. Which approach would be best: a) Let the project have a reference array to all of its reviews OR b) let the review have a projectId?
3. Moving all reviews to a new collection would mean a couple of millions documents in that collection. Would that cause slow queries? Such as this, which would be a common way of consuming the data: Reviews.find({projectId})