Drizzle vs. Prisma: why does Drizzle need less row reads?
I deployed a project to PlanetScale and noticed a high number of row reads for searching a table of 500 rows. I didn't think anything of it. I've been wanting to switch to Drizzle ORM anyway after I saw the TypeScript schema definition.
I made the switch with everything the same (except I'm not sorting by id) and noticed a HUGE rise in read/return efficiency with Drizzle. The SQL query is the same. The indices are the same. I did notice Prisma prepends the db name and table everywhere, like
I'm impressed. Is there anything I'm missing?
PlanetScale's row reads is generous, but there's a lot of savings here.
I made the switch with everything the same (except I'm not sorting by id) and noticed a HUGE rise in read/return efficiency with Drizzle. The SQL query is the same. The indices are the same. I did notice Prisma prepends the db name and table everywhere, like
db.Table.column wherever it can. Drizzle doesn't.I'm impressed. Is there anything I'm missing?
PlanetScale's row reads is generous, but there's a lot of savings here.
Solution
Yeah, prisma doest use joins, it makes individual calls and combines the data.
