drift
drift copied to clipboard
simolus3
[SQLParser] ATTACH DATABASE supported ?
Can you describe your use case for analyzing ATTACH statements? At the moment the analyzer doesn't check schema names at all (so something like SELECT * FROM foo.bar would fail). How much of the schema support do you need?
My future mobile app will have at less 5 databases.
Each database will have between 100 and 600 tables, each table between 10 and 600 fields, that's huge, but I can't make differently. So, sometimes, I will need to attach until 5 databases at the same time, and query them with join clauses.
How much of the schema support do you need?
Honestly, I don't know for the moment but I think I will want to use the maximum as possible of SQLite capabilities.
What you need is big data for mobile! Joke apart, I don't know your exact use case but I would first consider to review the data structure of your app and try to make something more scalable . If you cannot then I'd probably consider using sqflite directly without using a layer like moor. You will probably want the raw sqlite capabilities without adding more overhead.
Another use case which i am running into is having CMS exported database for offline first stored on a CDN, and a User database that possibly could be encrypted referencing the readonly database.
Example:
CMS DB (Read only): songs, albums, artwork
USER DB: playlists, likes, history
This would allow network updates of the CMS DB in the background and the user DB is created per account on device
I have done some testing with sqlite3 and can get it to work
The use case Rody described sounds pretty interesting, but it would require broader support for different schemas in core drift as well. I'm happy to have a discussion about the requirements and scope for that, but it should probably be done in a new issue.
sqlparser has supported the schema field in table references for a while now, but we don't (and likely won't) support ATTACH statements unless there is a community contribution for it - I don't need that parser functionality for drift files. So I'll close this one as being done, but we can continue the discussion about multi-schema support.