Possible further abstractions?

[Q16solver]

Hey! First of all, this looks amazing and definitely feels like abstracting on SQL is a definite win in my book. I was however reading on some points raised for example by author of Orchid Orm https://orchid-orm.netlify.app/guide. Especially on the problems of reusable code (possible solution raised was repositories though not sure how viable that is). Are there any plans/standards at all for abstracting at a higher level? For example I might want simple ways to create, update, read and delete entities and if there’s full type-safety in the bottom ts-sql layer, then this may be possible? Or is this completely left to the user as a user land problem?

[AndriiSherman]

Hi @Q16solver!

Yeah, currently we are just doing same things you can do with SQL. And as a user you are free to do anything you want. Map, abstract code, move code blocks in separate layers and so on. Our first priority - is not to block a user from full power of SQL and provide them as much flexibility as possible

But right after main core will be ready we were thinking on creating more abstract way to query your database, but it will be in different npm package and on top of existing drizzle-orm features

[Q16solver]

Sweet, that sounds amazing, a better abstraction layer that is built upon a solid foundation from sql that is fully type-safe can most likely prevent a lot of the weird edge cases when simply abstracting away from SQL which a lot of other ORMs like Prisma can run into, which spirals into thousands of github issues for implementing existing features that are supposed to already be in SQL. Personally I'm still using Prisma for it's simplicity in basic CRUD with relations that comes with great DX, but there are many gaps/overhead like the schema not being modularised etc that comes with developing a custom solution. Once the DX for drizzle becomes up-to-par with a regular ORM, I can see this being the de-facto specific domain solution to typescript projects