Document Versioning

Working with Data
,
#1

We are using C# driver in .netcore. We are looking for sound implementation of Document Versioning similar to SQL Server versioning. I have gone through BlogPost where they discuss document versioning.

Thanks

#2

Document versioning in MongoDB can be implemented in multiple ways, each with its pros and cons. Here are a few options:

  1. Embed version in the Document: Store a version field in each document and increment it with each update.
  2. Store each version as a separate document: Store each version of the document as a separate document in the collection, with a unique version identifier.
  3. Use MongoDB’s built-in change streams: MongoDB 4.0 and later support change streams, which can be used to track changes to a collection in real-time.
  4. Use a separate collection for version history: Store the version history of a document in a separate collection, which can be linked to the original document using a reference.

Whichever method you choose, make sure to test it thoroughly and to consider the trade-off between the complexity of the solution and the requirements for versioning in your use case.

#3

  1. Does the process which reads Change Stream have to be running consciously or it can have downtime? Is there the concept of “resume from” when I start reading from ChangeStream again? What are the provisions for consistency in this?

  2. I was also thinking about implementing versioning in my application itself with the below approach.

Below operations would be in the transaction

Insert for the document (newDocument) comes in with PrimaryKey = PolicytId

  • Get the CurrentVersion from the database using PolicyId
  • Assign the Version number of the newDocument as CurrentVersion + 1
  • Delete existing document and insert newDocument in collection
  • Also insert newDocument in history_collection or archive_collection

What do you think about this approach?

#4

Hi @Marmik_Shah

There are a couple of blog posts that may be useful to you regarding this topic:

Note that those two posts are quite technical, but contains a lot of interesting information regarding the pros/cons of some approaches. Those are a bit old, but still relevant today since it deals with schema design.

Best regards
Kevin

#5

Thanks for the links.

See