Query caching

It would be nice if queries would get out of the box support for caching, it’s a pain to make a wrapper since you’d have to make sure the function you create accepts and returns the right types. The syntax could be something like:

const users = await client.user.find({
  select: { x: true },
  where: { id },
  cache: true // defaults to y ms
})

// Custom time
const users = await client.user.find({
  select: { x: true },
  where: { id },
  cache: 60000 // cache for 1 minute / Maybe an options object for custom cache storage keys?
})

The cache could default to memory and a custom solution could be added in schema.prisma

datasource db {
  provider = "postgresql"
  url      = env("DATABASE_URL")
}

datasource cache {
  provider = "redis"
  url      = env("CACHE_URL")
}

Any thoughts? I just read about the middleware API which maybe could be used under the hood to support this feature, manually caching there would cause duplicate code

[AnandChowdhary]

I’d also love to see an easy way to invalidate the cache, if this is implemented. For example:

const users = await client.user.find({
  select: { x: true },
  where: { id },
  // Set a cache time of 1 minute and a key of "users-cache"
  cache: { time: 60000, key: "users-cache" }
})

Then, if you update a user, for example:

await client.user.update({ where: { id }, data: { /* new user data */ } });
// Invalidate the user cache
await client.cache.invalidate("users-cache");

This can probably even happen under the hood without explicit keys, because Prisma knows when user.update or user.delete are called, it can automatically invalidate any cache for user.find, etc. 🧠

[laurencefass]

+1 for official cache middleware support!