[SOLVED] Firebase. What's more efficient when using NoSQL – to store referenced table object in an object or to store id and retrieve it on demand?

Issue

So I am currently learning Firebase and NoSQL databases. I have a doubt about which to use when I have lots of referenced objects in other objects. I have come up with a solution to hold id of the referenced object instead of holding the whole object and retrieving it by id on demand but started wondering which way is more efficient and should be used.
First example is holding id of an object:

"job": {
    "id": 1,
    "name": "name",
    "orderId": 1
}

When using this method, whenever I get job object, I also get order object and attach it to the job.

Another solution is to hold whole object:

"job": {
  "id": 1,
  "name": "name",
  "order": {
    "id": 1,
    "date": "2022-01-28"
  }
}

In this way, there are no problems except that database has huge objects.

Which is better and more efficient to use? Or maybe there is a third way?

Solution

None of these solutions is better than another. You should use one, or the other according to your app’s use case. I’m not entirely sure if you are using Cloud Firestore or the Realtime Database, but you should always measure a structure versus the other. How can you do that? Simply by adding some dummy data and performing the needed queries. In this way, you’ll know which one is more expensive than the other.

Besides that, if you’re having a SQL background, duplicating data is normal with NoSQL databases. This technique is called denormalization. If you consider at some point in time to try using Cloud Firestore, then please check my answer from the following post:

• What is denormalization in Firebase Cloud Firestore?

Answered By – Alex Mamo

Answer Checked By – Mary Flores (BugsFixing Volunteer)