pg-mem VS @databases

I've used pgmem https://github.com/oguimbal/pg-mem for the last couple of years for the same thing.

I was wondering the other day how to classify tests that use a test double like pg-mem, which isn't a mock but isn't the Dockerized test DB either :

https://github.com/oguimbal/pg-mem

In my case, I use TypeORM with PostgreSQL, and there's pg-mem to run an instance in memory, it supports most of the common functionality of PostgreSQL but you will need to do some adjustment to your code to be within the limits.

Postgres in the browser is possible through pg-mem: "pg-mem is an experimental in-memory emulation of a postgres database" but it also suffers from no persistence. If you can persist to a file somewhere then read it in on startup (and if your local data isn't huge) this might work.

Dont get me wrong: I am trying to contribute to opensource as well, so I get that supporting small projects can be demanding. There's nothing wrong in not spending your weekends on OS. But not asking for help, nor specifying that a project is unmaintained, nor even answering issues & pull requests for years feels just wrong.

A pure Javascript in memory emulation of Posgres, to help writing better node tests https://github.com/oguimbal/pg-mem

Okay, I had a bit of spare time,I've implemented that, and it is now available with [email protected]

To get a glimpse of what I'm talking about, you can clone this repo and follow "Development" instructions (by the way this is a small OS lib I maintain, I wrote about it here)

At Databases because you don't need ORM.

Depending on the complexity of your queries, Prisma might indeed not be the best abstraction for you. If you're proficient in SQL and don't want to sacrifice type-safety, there are really nice, low-level alternatives to Prisma such as Zapatos, Slonik or atdatabases. We're laying this out in our docs here: Should you use Prisma?

I built https://www.atdatabases.org to make this as easy as possible to get right when querying SQL databases with node.js

Depends a lot on the size of your database records. We’ve found that for a few tables with big JSONB columns it can make a huge difference but for 90% of queries it makes very little difference. https://www.atdatabases.org with @databases/pg-typed or @databases/mysql-typed also keeps the types in sync with which columns you select.

Since the method is marked as async, you can use await to get the results of a query (if your database library supports promises. For example with https://www.atdatabases.org as your db library you could do

I found Prisma close but not quite there. That's part of what motivated me to keep working on https://www.atdatabases.org, which I think is already there as an enterprise ready ORM for node.js

I built https://www.atdatabases.org which has an ORM for node.js, but also supports writing SQL queries in a safe way. It is type safe, and has much simpler & more flexible transaction support than most node.js ORMs.

If you’re using TypeScript and don’t have untrusted user data, @databases can generate static types, which can be a good alternative to runtime validation.

Did either of you consider @databases? It has pretty much the same approach to SQL as Slonik. I’m curious if there’s any reason why Slonik is preferable?