Discussion regarding SQLite support and PR collaboration

[swoiow]

Hi @Maxime-J,

I’ve been following your discussion in issue #2578 and took a look at your project umami-sqlite. It's a great approach! I noticed you are using a 3.0.3.patch method to switch the database—very clever.

I've also been working on this and recently submitted a PR #4106 to modify the Docker configuration. I saw your PR #4126 as well.

Currently, I’m running Umami with PostgreSQL, but I strongly agree with you: for a lightweight analytics tool, SQLite is more than enough for most users. I’m keen to follow your progress and hope we can push for official SQLite support together. Let me know if there's anything we can sync up on!

Best regards.

[Maxime-J]

Hi, thanks for your message.

Structure will change a bit on next release, Umami having more and more features, a single patch file isn't quite adapted now.
Next version will require a customized SQLite, including regex operators and a specific stat function.

Full source in release and docker image will become the recommended ways to get it.

I won't push for official SQLite support though, while it does work really well, supporting multiple databases is time consuming.
For example a query optimized for PostgreSQL might not be optimized with SQLite.
It's way easier and more efficient to maintain and focus efforts on a single database.

Have a look here at the next release!

[swoiow]

Got it. Looking forward to the next release!