sql script to init DB + empty sqlite db initialised

[MagicalLay]

No description provided.

[But2ene]

ça m'a l'air fonctionnel,
une question sur les participations non notées, ça semble géré dans le code plutot que la BDD avec la structure actuelle

[Gbury]

It may be standard to have the IF NOT EXISTS, but since we have a DROP IF EXISTS above, aren't we guaranteed that the table does not exists at this point ?

[But2ene]

The redundancy appeared because I added the "DROP IF EXISTS" at the begininng.

Actually, the sql script is more like an executable specification documentation. This is a good design document.
Tables are created by the ocaml code, and I feel like it should be kept that way.

It might be interesting to automatically compare table definitions by ocaml with our sql specification some day.

avoiding if not exists in this case (because we are always deleting schema) ensure table definition is updated, but it becomes possible to accidently delete data if we execute the script on the production database. What do you prefer?

[MagicalLay]

to me there is no real point to have sql creation queries in 2 places: the lib and the script. it's just an opportunity for future desynchronisation. IMO, we should either:

• use the script to create the DB (a true option to me, I don't see what would be the true issue with that ; but I'm fine with other options as well), and make the lib only writes to it but not create the tables
• update the lib code to reflect on the new schema, but in such case we don't really need it outside (apart for work planification like now)

The only case it would make sense to me to keep it there alongside the lib code if the lib code is creating the table would be if it was written to by a Github action or sort of (like pgdump), to keep track of what the current database schema is ; a bit like a documentation. but in such case it would be just that, a documentation generated from the code, which is the real source of truth