TimescaleDB is actually an extension ontop of PostgreSQL. You can simply use the PostgreSQL connector. It is not too difficult to turn a Prisma generated table into a TimescaleDB hypertable. Here is the process:

Replace your PostgreSQL 10 docker image with: timescale/timescaledb:1.3.2-pg10

Then install TimescaleDB extension for your prisma database.

CREATE EXTENSION timescaledb;

Run a Prisma migration that creates the table that stores your time series data

datamodel.prisma

type EventTrainer {
  id: Int! @id
  user: User
  time: DateTime! @createdAt
  meta: Json
}

TimescaleDB hypertables PRIMARY KEYs must include all dimensions that you want to partition by. So you have to drop the PRIMARY KEY Prisma generated during the data model migration, and then create one that includes the "time" dimension.

ALTER TABLE "default$default"."EventTrainer" DROP CONSTRAINT "EventTrainer_pkey";

ALTER TABLE "default$default"."EventTrainer" ADD CONSTRAINT "EventTrainer_pkey" PRIMARY KEY("id", "time");

Then convert the table into a hypertable

SELECT * FROM create_hypertable('"default$default"."EventTrainer"', 'time');

I am able to insert and query data using Prisma client/bindings just fine with the above hypertable.
The @createdat decorator works and defaults to NOW() as expected.
Relation fields work

Nothing really seems to change from how you would normally use Prisma backed by vanilla PostgreSQL. Cheers my friends (-‿◦)

TimeScaleDB connector #1672

Description

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions