I'm wondering if the interactive transactions can be callable out of the callback scope like in the image:

...the idea is to use the this.data (which is the instance of prisma but from a service that is in the global scope via the Dependency Injection)

just for e.g. I can do something like this.data.request.findUnique(......) intead of using the prisma that comes in the callback, please let me know if that can be possible to continue getting the benefit of the transactions rollbacks if something is being throwed

Hey there, I was hoping someone might review my schema. I'm fairly new to Prisma. The goal is to allow a user to have their own projects that only they can access, users can also be a part of a group which will have projects that anyone in the group can access, and to allow there to be a manager user that can access multiple groups and their projects. Any feedback would be much appreciated!

generator client {
  provider = "prisma-client-js"
}

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

// -------------------------------------- //

enum Role {
  NONE
  GROUP
  MANAGER
}

enum ProjectStatus {
  TEMPLATE
  IN_PROGRESS
  COMPLETED
}

enum TaskStatus {
  IDLE
  DOING
  DONE
}

// -------------------------------------- //

model User {
  id    Int    @id @default(autoincrement())
  name  String
  email String @unique
  role  Role

  createdAt DateTime @default(now())
  updatedAt DateTime @updatedAt

  projects Project[]

  Group   Group? @relation(fields: [groupId], references: [id])
  groupId Int?
}

model Group {
  id    Int    @id @default(autoincrement())
  name  String
  email String @unique

  createdAt DateTime @default(now())
  updatedAt DateTime @updatedAt

  members  User[]
  projects Project[]

  Manager   Manager? @relation(fields: [managerId], references: [id])
  managerId Int?
}

model Manager {
  id     Int    @id @default(autoincrement())
  userId String @unique
  name   String
  email  String @unique

  createdAt DateTime @default(now())
  updatedAt DateTime @updatedAt

  groups Group[]
}

// -------------------------------------- //

model Project {
  id            Int           @id @default(autoincrement())
  createdAt     DateTime      @default(now())
  updatedAt     DateTime      @updatedAt
  projectStatus ProjectStatus
  archived      Boolean       @default(false)
  starred       Boolean       @default(false)
  name          String
  description   String
  tag           String
  deadline      DateTime
  links         String[]

  stories Story[]

  User    User?  @relation(fields: [userId], references: [id])
  userId  Int?

  Group   Group? @relation(fields: [groupId], references: [id])
  groupId Int?
}

model Story {
  id          Int      @id @default(autoincrement())
  createdAt   DateTime @default(now())
  updatedAt   DateTime @updatedAt
  description String

  tasks Task[]

  Project   Project? @relation(fields: [projectId], references: [id])
  projectId Int?
}

model Task {
  id        Int        @id @default(autoincrement())
  createdAt DateTime   @default(now())
  updatedAt DateTime   @updatedAt
  name      String
  status    TaskStatus

  Story   Story? @relation(fields: [storyId], references: [id])
  storyId Int?
}

Hey all 👋🏿 having an issue with a

findMany

query and

some

operator: When running this query:

const clusters = await prisma.cluster.findMany({
  where: {
    followers: { some: { id: userId } },
  },
});

I get this error:

ConnectorError(ConnectorError { user_facing_error: None, kind: RawDatabaseError { code: "unknown", message: "Command failed (Location40081): $in requires an array as a second argument, found: objectId)" } })

The two related models look like this:

model Cluster {
  id          String   @id @default(dbgenerated()) @map("_id") @db.ObjectId
  name        String
  description String?
  sparks      Spark[]  @relation(fields: [sparkIDs])
  sparkIDs    String[] @db.Array(ObjectId)
  user        User     @relation("ClusterOwner", fields: [ownerID], references: [id])
  ownerID     String   @db.ObjectId
  shared      Boolean  @default(false)
  deleted     Boolean  @default(false)
  followers   User[]   @relation("ClusterFollowers", fields: [followerIDs])
  followerIDs String[] @db.Array(ObjectId)
  legacyId    String?  @unique
  meta        Meta[]   @relation(fields: [metaIDs])
  metaIDs     String[] @db.Array(ObjectId)
  log         Log[]
  createdAt   DateTime @default(now())
  updatedAt   DateTime @updatedAt

  @@unique([ownerID, id])
  @@map("clusters")
}

model User {
  id               String         @id @default(dbgenerated()) @map("_id") @db.ObjectId
  firstName        String?
  lastName         String?
  name             String?
  email            String         @unique // TODO: doesnt seem to be being enforced
  type             UserType       @default(EXTERNAL_SEA_USER)
  pin              String         @default("0000")
  role             UserRole       @default(USER)
  deleted          Boolean        @default(false)
  legacyId         String?        @unique
  clusterFollowers Cluster[]      @relation("ClusterFollowers")
  clusterOwner     Cluster[]      @relation("ClusterOwner")
  processedForUser Log[]          @relation("ProcessedForUser")
  processedByUser  Log[]          @relation("ProcessedByUser")
  sparks           Spark[]
  subscriptions    Subscription[]
  createdAt        DateTime       @default(now())
  updatedAt        DateTime       @updatedAt

  @@map("users")
}

Any ideas on what I might be doing wrong or is this a bug with the mongo connector?

hello maybe someone can help me. I have an explicit Relation. I want to update the programs in activity everytime when i do an update on activity. At the Moment i get an Primary Constraint Fail. What can i do?

Hey all, I just had an interesting JS type problem removing nulls from Prisma queries which might be useful to someone: I'm using this function to remove nulls from my Prisma get queries:

export function removeNull(obj: any): any {
  if (!obj) return obj;
  if (obj instanceof Date) return obj;
  else {
    return Object.fromEntries(
      Object.entries(obj)
        .filter(([_, v]) => v != null)
        .map(([k, v]) => [k, v === Object(v) ? removeNull(v) : v])
    );
  }
}

And before I added the instanceof check it was removing all Date types from the response because JS types Date as an object 🤪 The result was returning objects with {insertedAt: {} ... } Maybe I'm missing a native Prisma method to remove nulls from queries??

If a Prisma query returns many rows it returns an object:

{0: {}, 1: {}, 2: {} etc...}

Has there been discussion about returning an array instead? An array would be more useful than an object whose keys are just the position in an array anyway. We can get the array with Object.values I know but still it seems like an unnecessary step..

How can I get autocomplete working with prisma in vscode using typescript?

How can I do a query where I'm saying NOT for a field with two values. For example, I want users without the name "Ethan" or "Jeff".

Hey, has this repository been deleted recently? github.com/prisma/prisma-client-lib-go Its a dependency of the generated go-client for prisma1. Fetching dependencies fails since Friday because of this prisma-client-lib-go.

I've looked up and down in the prisma docs and can't find if there is any API for getting the length of a scalar list, is there one?

Hey, curious how can I construct a query to fetch tag which has the most contents (relations)? Eg.

Tag {
    contents    Content[]
    name        String      @unique
}

Would need to order tag which has most content first.

Why is it that, when migrating my existing schema to modify a many-to-one relationship to a many-to-many, that I can't preform the creation of a table using

relation: { connect: [ { id: num }, ...]}

I found a bug in Prisma but I'm legit too lazy to reported

report it

Because I know y'all are going to want a whole ass report

In my schema I auto generate ids for this particular model

id          String        @id @default(cuid())

I have a special case where I need to generate the record using an id I generate. Is it possible to pass a custom id into

prisma.model.create

?

Hey everyone! I had a question about upgrading from Prisma 2.22.1 to the latest version (3.9.1). It seems when I run

migrate dev

after upgrading, it wants to create a new migration on existing SQLite and PostgreSQL databases. For SQLite, it wants to recreate all my tables (by creating a new migration basically with

-- RedefineTables

and

-- RedefineIndex

). For PostgreSQL, it does

-- DropForeignKey

,

-- AddForeignKey

, and

-- RenameIndex

(that I have seen so far). My main question is, is this expected and I should just add the new migration when upgrading (vs some other approach)? Any other tips/considerations to help migrate existing databases between these major versions? I did see this: https://www.prisma.io/docs/guides/upgrade-guides/upgrading-versions/upgrading-to-prisma-3 Thanks!

Tips for Content Creation by DevRel - Part 2 #shorts 📷 If you're looking to improve your content, or are considering starting to create content, follow these tips from a developer relations perspective! 👀 #shorts #prisma #tipsforcontent

Hi everyone! I have a general question about strategy. I have a nextjs app w/ my db on heroku. What I'm imagining I want is a staging DB that exists in order to test updates to my schema and data before actually going live. Have any of you done this? What would it look like to set this up? Is this a good idea? Is there a better idea? Any other thoughts? Thanks!

I'm considering replacing my graphql sqlserver implementation with Apollo and Prisma. I've got it basically working in NextJS, but I realize there is one step that is going to be a huge drag on the conversion and I'm wondering if there is a simple workaround like there is in EntityFramework which is what I'm converting from. Specifically, when. did the db pull, I got all my 100 or so tables like this:

model DiscussionItem {
  Id           Int       @id(map: "PK__Discussi__3214EC0771E234E0") @default(autoincrement())
  DiscussionId Int
  AttendeesId  Int
  DateEntered  DateTime  @db.DateTime
  MessageText  String    @db.NVarChar(4000)
  Attendees    Attendees @relation(fields: [AttendeesId], references: [Id], onUpdate: NoAction, map: "DiscussionItem_fk")
}

The real hassle is that all of the schema begins with uppercase "I" and I want those properties to translate easily to camelcase on the ApolloServer side. That is, id, attendeesId, discussionId, etc. The issue is the accepted norm for database columns is begin with capital, the accept norm for JavaScript variables is begin with lowercase. Is there an option for that?

@Peter Kellner You can use the @map and @@map attributes to map fields in your schema to the underlying names in the database. Essentially allowing you to pick whatever convention you like for the client while maintaining the names in the actual DB. https://www.prisma.io/docs/concepts/components/prisma-schema/names-in-underlying-database#mapping-collectiontable-and-fieldcolumn-names

Thanks @Sabin Adams I'm feeling a little slow here. Does this mean that the schema.prisma is always the same but some how you reference the variables in your client code makes them look different?

@Sabin Adams I think I get it now. @@map does the table name and just @map does the column name. That would mean I'd have to go slowly and carefully through my generated schema and for every column add a @map table? That sounds like it would be no fun, not to mention all the typo's I'd probably leave behind. I guess this means there is no way to have this happen automatically like it does in Microsoft's EntityFramework?

@Peter Kellner At the moment that is correct, there isn't an option to have that happen automatically. I had a similar problem with tons (50+) of tables on a large application I was working on. I ended up just fixing naming conventions as I needed specific tables and columns

So for example, I would introspect and leave it as is. Then in a PR where I accessed the

Users

table, I would fix the conventions in that model as part of the PR

Slowly everything has been getting moved over model by model and everything in the code follows those good conventions

E.G: If I needed to use the

User

model for the first time. It might look like

model users {
    Id Int @id @default(autoincrement())
    first_name String
    last_name String
}

I'd switch it to

model User {
    id Int @id @default(autoincrement()) @map("Id")
    firstName String @map("first_name")
    lastName  String @map("last_name")
    @@map("users")
}

Has anyone solved the missing

'_http_common'

issue? I see a lot of issues on GitHub (and mentions in search here), but haven't found a solution yet. Thanks. ☺️