Which database modelling practices should we follow?

Which is the best way to model our “database”?
Feels like NoSQL is the best way but I might be wrong.
Should we be using denormalisation for example? Or are there limitations that will bite us in the long term?

2 Likes

There have been a few discussions on the topic. Please see the following:

I think it will depend on your application and how you want to handle data relationships. You can experiment with a SQLite canister posted herehttps://forum.dfinity.org/t/database-like-searching-querying-with-bigmap-bigsearch/1977/25?u=hassen. You can wait for the eminent release of a graphql canister by @lastmjs. This will allow you to experiment and get an early feel of what each model offers. I am sure many more options are going to be available in the future.