Here's my shameless plug: `pip install sqlitedao`: https://github.com/Aperocky/sqlitedao
Here's my shameless plug: `pip install sqlitedao`: https://github.com/Aperocky/sqlitedao
my experience optimizing SQL has been removing the ORM layer via hand rolling SQL queries.
Most of my peers also seem to agree that ORM is great for knocking a feature out, but you’ll inevitably run into performance concerns at scale and need to write your own optimized queries yourself.
I feel like with SQLite having limited resources, you’d likely want to optimize your queries as best you can without an ORM? Is that not the case?
The fundamental issue with both handrolled SQL and ORM is inefficient queries leading to linearly increasing time consumption. for instance usage of `LIMIT` and accidentally querying on non-indexed fields. But I don't see how this differentiate the 2 as I have seen it happen in both situations professionally.
A good ORM enforces certain behaviors, specifically querying by index and constant time pagination. This guards against accidental querying behaviors.
ORM's never scale and you end up with hand written SQL long term. Unless it is really just a CRUD app then ORM's really work fairly well.