ml-dotnet-client-api | A .NET client for the MarkLogic NoSQL database | Document Database library

When you set a key's value to null in Fauna, it is removed from the object. In your example, assuming ref is a valid Reference:

UPDATE based on comments:

Updated query:

You can do (almost!) anything with an aggregate query. In your case I suggest using $unwind to convert the lists to onjects, then $match on your target field(s), $project to trim down the output, $replaceRoot to simplify the structure and $limit for good measure as there's actually 2 records that match your criteria.

https://mongoplayground.net/p/UsKeqA0aWYK

When you are in a transaction with read concern snapshot, you are reading from a consistent snapshot of the data.

Otherwise you could be experiencing various phenomena described here.

You should probably change

Using Azure CosmosDB for the storage could help. With CosmosDB, you can easily store the image as an attachment to the student document: https://docs.microsoft.com/en-us/rest/api/cosmos-db/attachments

If you don't plan to use CosmosDB, be more precise please about the database storage.

You are mostly correct.

Firstly MongoDB is unstructured by nature, hence i recommend not trying to find "analogies" to match it to the structured scheme. With that said similarities do exist so for simplicity we can do so.

One more note is that this syntax your referring to is mongoose specifically and not the original Mongo syntax.

1. Mongoose.schema( { } ) Yes, mongoose gives you the ability to "force" a structure, note this will only come into play when trying to insert/create/update documents and not when it comes to querying.

2. Mongoose.model('Name', Schema) Yes-ish, mongoose will not create a database per-se, meaning if it does not exist a new one will be created, however inserting a document to that model will create a such collection.

3. new Model ({ //DATA }) Yes, however you need to add new Model().save(), without the save it will not be triggered and saved into the database.

4. new Model ().query() Yes-ish, again similar to the model function this is a mongoose wrapper (that I'm less familiar with) As specified in the docs:

Query constructor used for building queries. You do not need to instantiate a Query directly. Instead use Model functions like Model.find().

Personally I just use the Model functions to query such as find,findOne,aggregate and more.