Bom dia galera, tudo bem? É o seguinte sempre trabalhei com ORM no backend NodeJS, antigamente eu utilizava o Sequelize e depois comecei a usar o TypeORM, mas de uns dias para cá comecei a estudar o PRISMA, até tenho gostado dele, mas estou com a seguinte dúvida. Como faço para reverter a última Migrate com ele, no Sequelize e TypeORM isso funciona sem probleas, mas no Prisma não consegui, até encontrei alguns comandos para zerar o banco, mas eu queria apenas a última migrate. Encontrei o seguinte comando em minhas pesquinas prisma migrate down --experimental no meu caso eu coloco: yarn prisma migrate down --experimental, mas não funciona. A versão que estou usando é a 3.10 Humildimente pergunto se alguém trabalha com ele e poderia me orientar? Abraço a todos...

Hello! Is there a way to filter a subselection? Like if you have

prisma.authors.findMany({
   where: {
       books: {
          id: {
              IN: [1,2,3]
          }
     }
   }
})

This would filter the authors, but what i want is to get all the authors, but filter their book selection. How would I do that?

Hello everyone! I discover Prisma as a student. And I'm having some difficulties. I can't get it to work so that when I submit a user with an email address already in the DB, it returns something other than a "unique field violation" error. Please, if someone could help me 😞 thanks

👋 Hello, team!

can we do GridFs using prisma

Mongo db storage

Prisma timing out on Heroku Postgre is so frequent I'm tired. This just failed on my bitbucket-pipeline as it was deploying to Heroku. I will have to remove the postinstall script from my package.json.that runs "npx prisma migrate deploy" , because I don't want my deploys to fail. Then I will have to resort to running "migrate deploy" locally, copy all DATABASE_URL of all my environments being deployed to, add them to my .env file . And manually run "migrate deploy" for all of them. Is there any reason these things happen frequently ?

Anyone ever gotten a Postgres Bytes field to work with $executeRaw()? Something is trying to base64 encode whatever is passed in in a way that is making it impossible for me to actually store the bytes I want in the database. Buffers work correctly in generated executors, but I am trying to use executeRaw to combine some operations that need to be in a transaction.

Hi everyone, Anyone have any idea how to handle the best way revisions of a content in prisma with PostgreSQL?

Hello Any plans on supporting

tstzrange

type?

Hi, I have 2 questions: 1 - Is following a good way to structure logic? LOGIC => A table for teams. There can only be 2 users in a team.

model User {
  id              String               @id @default(cuid())
  team            Team[]
  ...
}

model Team {
  id          String        @id @default(cuid())
  member      User[]
  ...
}

QUESTION: Is this logic good or would it better to have two fields (member1 & member2) within the

Team

table, which would reference

User

? 2 - The above logic is incorporated within Prisma. How do I query

Team

to get

User

. I get stuck because there are no foreign keys within Team referencing User.id, nor is there another table connecting Team -> User. (correction:) - There is a table

_TeamToUser

which only has 2 fields (A, B). (correction:) - Got this working with the following query.

select * from "Team" inner join "_TeamToUser" on "Team".id = "_TeamToUser"."A" inner join "User" on "_TeamToUser"."B" = "User".id;

Would love feedback to 1st question.

I have an old Prisma 1 Cloud project that I need some help with. I will be later this year moving it over to Prisma 2, but my client is needing an audit of the current system. What I need is some info about Prisma 1 Cloud’s architecture but am struggling to find the relevant documentation. I need some clarity on how the db hosting is done (responsibilities, hardening, up-to-date software, encryption of the data). Can anyone point me in the right direction?

Hi there, I'm thinking about utilising prisma in an up coming project. Could someone sell me the virtues of using Prisma over say, Firebase, as an ORM of choice.

The stack is Go and Typescript, so it has come well recommended 🙂

how was Prisma made. I've been looking at the codebase & Wondering where you get the knowledge to be able to build such a beast

Hi guys ! I have few random errors:

P1001 Can't reach database server at xxxx:xxx

So the Prisma client seems to lose the connection with the Postgresql DB obviously ... It’s not linked to the server load or too many connections because I have the same issue in staging environment (low trafic) It’s happening on two different services: a NodeJS API and a web worker Everything looks fine on Postgresql DB side and I have another very similar service connected to the same Postgresql DB but with Sequelize as ORM and no issue in this service Do you have any idea, what should I look for in order to fix it ?

Any idea why prisma might be spending so much time in

readdirSync

?

Hi, how can I filter all my customers who have a birthday today?

const customers = await prisma.customer.findMany({
   where: {
     birthday: {
       equals: 'YYYY-03-11',
     },
   },
});

Hi, I'm very new to prisma so please forgive me if the question sounds basic. How can I get the types of some particular Model to work under typescript. I'm not inferring them from a prisma query or mutation but I need to use them to work with something else. I came up that Prisma.<model>Select seems to give me the types of the Model I need. Am I correct?

Hello. Having some trouble inserting a record using

executeRaw

, whenever I try to execute this code snippet I get the following error

db error: ERROR: invalid input syntax for type timestamp with time zone: "$1"

const { from, to } = createAppointmentDto;
await prisma.$executeRaw`insert into public."Appointment" ("date", "patientId", "itemId") values ('[${from}, ${to})', ${id}, ${itemId});`;

The weird part is that if I hardcode the dates directly as strings, IT WORKS!

await prisma.$executeRaw`insert into public."Appointment" ("date", "patientId", "itemId") values ('[2010-01-01 14:45, 2010-01-01 15:45)', ${id}, ${itemId});`;

Surely I'm missing something but don't know what it is. Thoughts?

Prisma Meetup - Chance Strickland - Update While You Wait: Optimistic UI with Remix and Prisma ◭ Building UIs with instant user feedback when interacting with remote data has always been a challenge. We’ll show how Remix provides powerful APIs that make it easier than ever to perform optimistic updates, while Prisma beautifully simplifies the interface for updating your data. ◭ Chance is a software engineer at Remix who loves traveling, surfing, and helping developers like you build better websites and apps. ◭ Get in touch with Chance: https://twitter.com/chancethedev ◭ Join our Prisma Meetup group: https://www.meetup.com/pro/prisma-meetup/

Is it possible to filter on include with a parent value to the include? Assuming I want to take the post that were created by user , 2 days after signing up . In the includes I need to write a condition for the created_at of post to be gte to the user's created_at

Hello good people! I'm building a football scouting app, actually rewriting it to nestjs & prisma. I have a following set of tables (among many others) - PLAYERS, TEAMS and TEAM_AFFILIATIONS:

model TeamAffiliation {
  id        String    @id @default(cuid())
  startDate DateTime
  endDate   DateTime?
  createdAt DateTime  @default(now())
  updatedAt DateTime  @updatedAt

  // Relation fields
  player   Player @relation(fields: [playerId], references: [id])
  playerId String
  team     Team   @relation(fields: [teamId], references: [id])
  teamId   String
}

This table's responsibility is to keep the information to which team has the player been assigned within a certain time period (between

startDate

and

endDate

). I also have a NOTES table, its simplified model looks like this:

model Note {
  id               String   @id @default(cuid())
  docNumber        Int      @default(autoincrement())
  description      String?
  createdAt        DateTime @default(now())
  updatedAt        DateTime @updatedAt

  // Relation fields
  player           Player?         @relation(fields: [playerId], references: [id])
  playerId         String?
}

And here comes my question. On a note itself (when I'm for example fetching a list of notes) I need information to which team has the player been assigned when the note was created (it'd be perfect if I could filter notes by this piece of data). How would you approach this kind of problem? The simplest way would be just to create a field like

playerCurrentTeam

on a note and save it when a note is created. But it seems like a workaround, is there a better solution?

Hi! Kindly asking someones help. I need to insert a

work

record which has link to

order

,

user

and

location

. Schema is as follows:

model User {
  id            String          @id @default(cuid())
  email         String?         @unique
  orders        Order[]
}

model Order {
  id                 String          @id @default(cuid())
  orderNumber        Int
  user               User            @relation(fields: [userId], references: [id])
  userId             String
  origin             Location        @relation("OrderOrigin", fields: [orderOriginId], references: [id])
  orderOriginId      String
  destination        Location?       @relation("OrderDestination", fields: [orderDestinationId], references: [id])
  orderDestinationId String?
  work               Work?           @relation("OrderWork")
  includeLocations   Location[]      @relation("OrderIncludeLocations")
  excludeLocations   Location[]      @relation("OrderExcludeLocations")
}

model Work {
  id                    String     @id @default(cuid())
  origin                Location   @relation("WorkOrigin", fields: [workOriginId], references: [id])
  destination           Location   @relation("WorkDestination", fields: [workDestinationId], references: [id])
  stops                 Location[] @relation("WorkStops")
  originLocationId      String
  destinationLocationId String
  workOriginId          String
  workDestinationId     String
  workStopsId           String?
  order                 Order      @relation("OrderWork", fields: [orderId], references: [id])
  orderId               String     @unique
  status                WorkStatus @default(value: BOOKED)
}

model Location {
  id                    String  @id @default(cuid())
  address               String
  city                  String
  state                 String
  locationCode          String?
  workStops             Work[]  @relation("WorkStops")
  workOrigin            Work[]  @relation("WorkOrigin")
  workDestination       Work[]  @relation("WorkDestination")
  orderOrigin           Order[] @relation("OrderOrigin")
  orderDestination      Order[] @relation("OrderDestination")
  orderIncludeLocations Order[] @relation("OrderIncludeLocations")
  orderExcludeLocations Order[] @relation("OrderExcludeLocations")
}

Now when

work

is created

order

status becomes

BOOKED

. I can’t figure out how to upsert locations.

let work = await prisma.work.create({
    data: {
        destination: {
            connectOrCreate: {
                where: {
                    // address: data.work.destination.Address,
                    // city: data.work.destination.City,
                    state: data.work.destination.State,
                },
                create: {
                    address: data.work.destination.Address,
                    city: data.work.destination.City,
                    state: data.work.destination.State,
                },
            },
        },
        origin: {
            connectOrCreate: {
                where: {
                    // address: data.work.origin.Address,
                    // city: data.work.origin.City,
                    state: data.work.origin.State
                },
                create: {
                    address: data.work.origin.Address,
                    city: data.work.origin.City,
                    state: data.work.origin.State
                },
            },
        },
        // stops: data.work.stops.map(stop => {
        //     return {
        //         connectOrCreate: {
        //             address: stop.Address,
        //             city: stop.City,
        //             state: stop.State,
        //         },
        stopsCount: 0,
        workOriginId: data.work.workOriginId,
        originLocationId: data.work.originLocationId,
        workDestinationId: data.work.workDestinationId,
        destinationLocationId: data.work.destinationLocationId,
        workStopsId:"1234",
        status: data.work.status,
        orderId: orderId,
    }

Hey everyone! Just popped in to ask about the status of Prisma's Data Proxy. Currently it requires aliasing using

esbuild

or similar when used with Cloudflare Workers (otherwise it gives a "not allowed in browser" error). What's the blocker on removing this bundle requirement?

Hola a todos espero que esten muy bien, por casualidad alguien deployo prisma en heroku?

any updates on https://github.com/prisma/prisma/issues/5837?

Hey everyone! I could not find anything about indexing the referencing side's key(s) in relations with PostgreSQL in the Prisma docs. As that is something PostgreSQL doesn't automatically do (and neither does Prisma, see #3943), but often leads to huge performance improvements for large joins, I would suggest adding a corresponding advice to the docs (for example on the relations page). I would appreciate it, as I think many developers are not aware of this (multiple other ORMs create these indexes by default for the same reason). What do you think?

Hey everyone! I'm working on Tweetscape and am trying to move over to using Prisma, but I'm not sure if your nested querying capabilities will support something like:

select
      links.*,
      clusters.id as cluster_id,
      clusters.name as cluster_name,
      clusters.slug as cluster_slug,
      sum(tweets.insider_score) as insider_score,
      sum(tweets.attention_score) as attention_score,
      json_agg(tweets.*) as tweets
    from links
      inner join (
        select distinct on (urls.link_id, tweets.author_id, tweets.cluster_id)
          urls.link_id as link_id,
          tweets.*
        from urls
          inner join (
            select 
              tweets.*,
              scores.cluster_id as cluster_id,
              scores.insider_score as insider_score,
              scores.attention_score as attention_score,
              to_json(influencers.*) as author,
              to_json(scores.*) as score
            from tweets
              inner join influencers on influencers.id = tweets.author_id
              inner join scores on scores.influencer_id = influencers.id
            ${filter === 'hide_retweets' ? `where not exists (select 1 from refs where refs.referencer_tweet_id = tweets.id and refs.type = 'retweeted')` : ''}
          ) as tweets on tweets.id = urls.tweet_id
      ) as tweets on tweets.link_id = links.id
      inner join clusters on clusters.id = tweets.cluster_id
    where clusters.slug = '${params.cluster}' and expanded_url !~ '^https?:\\/\\/twitter\\.com'
    group by links.id, clusters.id
    order by ${sort === 'tweets_count' ? 'count(tweets)' : sort} desc
    limit 20;

While there definitely is a way to get all of that nested data, I can't figure out how to aggregate and sort based on the nested select data:

const links = await db.links.findMany({
    select: { // There's no way to `_sum` the nested fields from `scores`.
      urls: {
        select: {
          tweets: { // There's no way to have a nested `groupBy` here.
            select: {
              influencers: {
                select: {
                  scores: {
                    where: {
                      cluster_id: params.cluster,
                    },
                  },
                },
              },
            },
          },
        },
      },
    },
  });