cornucopia VS diesel_async

There are multiple queries each separated by ; and on top of each query, there's a comment giving a name to the query (it's more like a header)

I think the only thing that would require specific support in postgres_lsp is using the :parameter_name syntax for prepared statements [1] (in vanilla Postgres would be something like $1 or $2, but in Cornucopia it is named to aid readability). But, if postgres_lsp is forgiging enough to not choke on that, then it seems completely fit for this use case.

[0] https://github.com/cornucopia-rs/cornucopia

[1] https://cornucopia-rs.netlify.app/book/writing_queries/writi...

Some examples for anyone else reading:

https://github.com/kyleconroy/sqlc

https://github.com/cornucopia-rs/cornucopia

This is my preferred method of interacting with databases now.

Very flexible.

I like Cornucopia. It’s a SQL-first approach, so I don’t have to worry about an ORM generating pathological queries. It’s also basically zero cost compared to directly using rust-postgres and supports both sync and async. I also like that my SQL queries end up separate from my Rust code, so it’s easy to update all the relevant queries when the schema changes.

Also, I moved away from SQLx due to slow compile times and now use https://github.com/cornucopia-rs/cornucopia

While we're on the subject of ORM's I really like the https://github.com/cornucopia-rs/cornucopia way of doing things.

Basically write SQL in a file and code generate a function that runs the SQL for you and puts it into a struct (this one is for rust)

I think there's a library to do the same thing with typescript.

For me, the best way to talk to the database is with SQL and I don't have to learn an ORMs way of doing it.

You can take a look at https://github.com/cornucopia-rs/cornucopia which is a thin codegen layer on top of tokio-postgres for ease of use.

I think we have that https://github.com/cornucopia-rs/cornucopia

The best solution I've ever seen is this Rust library https://github.com/cornucopia-rs/cornucopia

You write plain SQL for you schema (just a schema.sql is enough) and plain SQL functions for your queries. Then it generates Rust types and Rust functions from from that. If you don't use Rust, maybe there's a library like that for your favorite language.

Optionally, pair it with https://github.com/bikeshedder/tusker or https://github.com/blainehansen/postgres_migrator (both are based off https://github.com/djrobstep/migra) to generate migrations by diffing your schema.sql files, and https://github.com/rust-db/refinery to perform those migrations.

Now, if you have simple crud needs, you should probably use https://postgrest.org/en/stable/ and not an ORM. There are packages like https://www.npmjs.com/package/@supabase/postgrest-js (for JS / typescript) and probably for other languages too.

If you insist on an ORM, the best of the bunch is prisma https://www.prisma.io/ - outside of the typescript/javascript ecosystem it has ports for some other languages (with varying degrees of completion), the one I know about is the Rust one https://prisma.brendonovich.dev/introduction

Does Cornucopia satisfy this requirement?

So I want to use diesel_async in my web backend.

The already linked discussion from last time already contains a lot of information about this. Diesel itself does not provide async operations and that will likely remain that way for a foreseeable future. At least my preferred solution is to keep async support in a separate crate. A prototype for this is currently available here. Keep in mind that this is not released yet, so there might be bugs everywhere. I plan to cut a first release of this crate after the final release of diesel 2.0, which means hopefully soon. As for ETA's: I generally do not give any ETA's for releases, as this is currently a free time project for me.

So, diesel is an ORM that tries to take full advantage of rust's typing expressivity to allow for statically checked, and fast, queries. I absolutely loved it when trying it out the first time.

i don't see why you or anyone else would consider it too big of an issue that Diesel doesn't have async. For those who really want async diesel right now, the author already released diesel_async as a stop-gap solution, but even without that there's nothing wrong with using spawn_blocking. It feels worse than it really is to use blocking thread pools; until io_uring is a thing, there's no real getting around the necessity of threads being blocked for I/O and so adding async to the mix doesn't magically make things faster.

As for now there is an prototype available here. It's technically feature complete, but depends on a unpublished diesel version + has some remaining bugs with certain mysql versions. If that is fixed I will likely publish a first alpha version officially. That can take same time because that's a second large project that needs maintenance time beside diesel itself and that's quite a lot to do in my free time. You can support this work by sponsoring me on github

Async support for diesel currently lives in a separate repository as there are language level blocking issues for publishing a version of this crate where we could commit to a stable release at all. See the corresponding diesel issue for details.

https://github.com/weiznich/diesel_async tho'.

Weiznich (the maintainer of Diesel) has created an experimental async version of the diesel Connection and RunQueryDsl traits, which should help with ease-of-use for Diesel within async contexts. It is not yet published on crates, but you can find it here: https://github.com/weiznich/diesel_async