Just playing with Interactive Transactions preview today and noticed a missing type shortcut that's really needed:

export type PrismaTransactionClient = Omit<PrismaClient, '$connect' | '$disconnect' | '$on' | '$transaction' | '$use'>;

Can this be added to the client generator?

Looks like there's an issue for this: https://github.com/prisma/prisma/issues/12436

Hi! I'm running a NEXTjs with Prisma/Postgres db Wondering if someone can take a look at this gist to help me out. I'm having trouble with me model relations. https://gist.github.com/sebass-surfer/3a300a0d50abdcdad9bffbe9a05084c2

Hi guys! How do I remove the square brackets in there? I want it to look like a json object.

Hi! Is there a way to disconnect multiple records (but not all)? Been looking through the docs and

updateMany

seems like the closest thing, but it’s only letting me update scalar fields

Hey guys queryraw doesn't work for me, data appear on prisma studio and there is no problem but when I try to run the query: const users = await prisma.$queryRaw`SELECT * FROM User`; , It only shows my postgres username. When I check my data on psql, It shows the same, tables appear but there is no data inside.

I'm trying to follow the examples to perform an upsert operation with a nested

connectOrCreate

call. The error I'm getting is saying that the

author

arg is unknown. Specifically:

Unknown arg

author

in create.author for type TweetUncheckedCreateInput. Did you mean

authorId

? Available args:

type TweetUncheckedCreateInput {

id: String

authorId: String

createdAt: DateTime

entities?: Json | Null

media?: MediaUncheckedCreateNestedManyWithoutTweetsInput

mediaKeys?: TweetCreatemediaKeysInput | List<String>

publicMetrics: Json

text: String

type?: String | Null

}

The relevant client code is here: https://gist.github.com/marr/c574e7a854223e828e652753d08dd117#file-twitter-ts-L42 and relevant schema is here: https://gist.github.com/marr/0b2e6a58858388df6b448f5d1d2dd879#file-schema-prisma-L23 Can anyone help me troubleshoot this?

Would the nested models be an issue on mongo or something? I'm thinking I just remove the relations and manually query with an aggregation pipeline

Do you use Prisma with AWS CDK? Have you wanted a really amazing prisma lambda layer with just the minimum needed to run prisma and the proper node

externalModules

settings? Look no further! This was developed over many days of grueling testing and trial and error. I hope to spare others the same headache.

Would a column named

type

cause problems?

How would I receive a string from my datetime column queries?

I'm trying to iterate over a result, but getting an error:

Type 'JsonObject' must have a '[Symbol.iterator]()' method that returns an iterator

Code in question (mongodb):

const res = await db.tweet.aggregateRaw({

pipeline

});

for (const tweet of res);

Is there any substantial difference in using

connect

or not when creating a row with a foreign key field? I note the result is the same but what would I choose one over the another one? Or this just for a declarative/semantical reason? Example:

model User {
  id            String    @id @default(cuid())
  name          String?
  email         String?   @unique
  emailVerified DateTime?
  image         String?
  accounts      Account[]
  sessions      Session[]
  rooms         Room[]
}


model Room {
  id          Int    @id @default(autoincrement())
  userId      String
  title       String
  description String @db.Text
  user        User   @relation(fields: [userId], references: [id])

  @@unique([userId, title])
}

From the above schema, If I would like to add a new room I could do: Let's say

user

hs already being resolved to a user found in the User model

prisma.room.create({  data: { title, description, user: { connect: { id: user!.id } } }})

or also,

prisma.room.create({  data: { title, description, userId: user!.id }} )

What is the correct way?

If I run SELECT ... FOR UPDATE inside an interactive transaction, when is the row actually locked? Is it when that line of code runs, or is it when the transaction is committed?

Is it possible to create a field that is unique and can be optional ? Like in the example?

phone        String?       @unique

For MongoDB

Not sure how many times this has been asked before, but is there a preferred way of accessing custom claim from an access token inside a @Query()? I have a custom claim that has an array if strings I'd like to use in a query

where: { id: { in: identifiers } }

There is so much inception going on that I have hard time keeping up... Express -> NestJS -> Prisma

👋 Hey folks. Quick Typescript question - how do I access nested fields when using the auto-generated Typescript interfaces from Prisma Client? For example, the model

User

contains a 1-many link to

Services

. We import the

User

interface but it only has

serviceId

, so in our code

user.service

is generating a typescript error (even though the object is present in the data returned from the DB). Sometimes we access objects several levels deep (i.e.

user.service.booking

), just wondering if there's a recommended approach for this.

Invalid

prisma.queryRaw()

invocation: Your raw query had an incorrect number of parameters. Expected:

0

, actual:

1

. What am I doing wrong here?

I'm trying to filter by null values, but am getting the following error:

Type '"AnyNull"' is not assignable to type 'JsonNullableFilter | undefined'.

My client configuration looks like this:

generator client {

provider        = "prisma-client-js"

previewFeatures = ["filterJson", "interactiveTransactions"]

}

I have a postgresql database, my prisma client version is 3.5.0, and the field I am trying to filter by is an optional JSON field. the findMany call:

await prisma.chart.findMany({

where: {

oldState: Prisma.AnyNull

},

});

What’s the best way or suggested approach to map prisma data models to graphql types? I’m considering having pure functions acting as mappers that are testable, but first curious if there is something that I can already use. Thanks in advance

Hello everyone! I am having a problem, I have a dockerized application and I get the following error:

I'm using lambda and I got this error

FATAL: sorry, too many clients already

I have

connection_limit=5

on my

DATABASE_URL

Any idea what to do about this specific error? I think it means too many connections for the specific client?

Can anybody explain or point me to documentation regarding unchecked vs regular input args?

Is there a way to configure prisma to create date columns (createdOn, updatedOn) on models defined in the schema without having to explicitly create them?

Hello everyone! How do you do the prism generate inside your Dockerfile? I do it with

RUN npx prism generate

but it makes my image too big.

Hey all, I'm trying to query a Postgres JSONB field (and avoiding using raw) with the js client. Filtering by an array's object isn't supported w/ the Postgres connector yet - only MySQL as of right now. Performance isn't really a huge deal for me since the query is run once per day and isn't user-facing, but does anyone know if there's a (sane) way to use Prisma to return values from the JSONB column defined below where

name

is equal to

foo

? My current hack in raw SQL for when I don't feel like parsing and filtering like an adult is a shameful display of hubris with

items::TEXT ~ 'foo'

lol. Data shape in the "items" column:

[
  {
    "name": "foo",
    "value": "example_1"
  },
  {
    "name": "bar",
    "value": "example_2"
  },
]

I'm currently using

prisma.$on('query', event => {});

to log queries, but I'd like to be able to associate these queries with a particular HTTP request In the rest of my codebase, I have a unique "request ID" that I can pass around, so that any logs from the same request can be associated with each other Is there a way for me to achieve this with regards to logging what Prisma is doing please?

I just took a look because I was curious, too. The generated code is pretty strict about what you can pass w/ your query config.. I have a hacky idea if you want to try that lol

you can add a debug string to all of your models and write a custom middleware to remove it from the params (where, data, etc) before sending it to the db