Hmm, was there some breaking change to how template strings are implemented for

queryRaw

in Prisma 2.23.0?

await prisma.$queryRaw(`
    SELECT c."uuid", c."createdAt"
    FROM "EmploymentRecord" AS er
    JOIN "User" AS u ON (u."uuid" = er."userUuid")
    JOIN "Company" AS c ON (c."uuid" = er."companyUuid")
    WHERE er."endDate" IS NULL
    AND LOWER(u."email") IN (${Array.from(prodUserEmails)
      .map((email) => `'${email}'`)
      .join(', ')})
    GROUP BY c."uuid", c."createdAt"
    ORDER BY c."createdAt" ASC;
  `);

The query above works perfectly and returns my several hundred rows, but:

await prisma.$queryRaw`
    SELECT c."uuid", c."createdAt"
    FROM "EmploymentRecord" AS er
    JOIN "User" AS u ON (u."uuid" = er."userUuid")
    JOIN "Company" AS c ON (c."uuid" = er."companyUuid")
    WHERE er."endDate" IS NULL
    AND LOWER(u."email") IN (${Array.from(prodUserEmails)
      .map((email) => `'${email}'`)
      .join(', ')})
    GROUP BY c."uuid", c."createdAt"
    ORDER BY c."createdAt" ASC;
  `;

Does not. I know I am manually string quoting my values, but not sure how else to perform an

IN

with

queryRaw

Hi all-- I'm working on using

prisma migrate deploy

with a Vercel deployment workflow + digital ocean pooled postgresql db & pgbouncer and can't quite figure out the right way to define a different DATABASE_URL to run during build time vs. runtime (in order to point to the non-pgbouncer URL during migrations at build time). It looks like Vercel exposes the "CI" env variable during build time which I thought would be helpful but is there a way to include a conditional in the schema.prisma file to react to this? Something like the following (which does not work for obvious reasons)?

datasource db {
  provider = "postgresql"
  url      = env("CI") ? env("DATABASE_URL_NOPGBOUNCER") : env("DATABASE_URL")
}

Thanks!

I think I'm missing something about how migrations are supposed to be used. Specifically I don't understand why to update my schema during development I need to point prisma to a running database instance. I get that

migrate deploy

needs to point to an actual database instance but why does

migrate dev

? To do its work it just needs the schema and the list of previous migrations, no?

is there any limitation for

notIn

operator? for example we have 5 users, 3 is in the condition and 2 is not. if we execute

prisma.user.findMany({where: {notIn: [...]}})

will return 7 result. the above is to clarify the problem and the real count may be 5000 items in

notIn

clause which returns wrong result. But

prisma.user.count(...)

and

prisma.user.findMany({where: {NOT: {in: []}}})

are ok

I have logged the sql for query like above. prisma will split

notIn

clause to 5000 per chunk and merge each result which result in the wrong total count results

For

In

cluase, the merge of each chunk is ok. And for

{NOT: { in: [] } }

it will not split params to chunks. So both of this situation are ok.

const v = await prisma.user.findMany({
    where: {id: {notIn: Array.from({length: 6000}, (v, idx) => idx.toString())}}
  })
  const total = await prisma.user.findMany({
    where: {}
  })
  console.log(v.length, total)

this should reproduct the problem.

Hey! I am currently doing something like

const lastLesson = await db.lesson.findFirst({ where: { date: { gt: new Date() } }});

but I have a field called

lesson.duration

. So I would ideally want to do only get it if the duration has passed, maybe something like this

const lastLesson = await db.lesson.findFirst({ where: { date: { gt: addMinutes(new Date(), 
this.duration) } }});

basically can I get the where to know context about the current row somehow so it can know what the duration is while checking, rather than having to loop? Is this possible?

hey! is there a recommended way how to know which versions of nexus and prisma work together? Because I guess I can't update to the latest prisma version right?

Is it possible to set those query param options outside connection URL? For example as separate variable in prisma.schema file?

Can I use relations when I have two different schema files in two different projects?

Hey! Second question in a day! Has anybody figured out how to keep Algolia and prisma/postgres in sync? I tried writing some middleware a while ago which was like https://gist.github.com/Swapnull/9a88e7b9df1db7a68711728394717355 , but it feels pretty crappy having to write a function for each model I want to put there. And that doesn’t even touch things like updating.

How do Prisma users keep track of functions in their DB?

Hi there. We are still using Prisma 1. I want to migrate to Prisma 2. I would like to know if there are performance benefits. What do you thing?

hey prisma guys, is there a feature request open anywhere about typing Json fields, like maybe referencing a typescript file and type within it with an annotation

Hey everyone, so currently I'm trying to migrate our development database, and I'm getting an error every time I try to do so. I've retried multiple times, nothing seems to work. Any help?

When and how to run prisma migrate deploy? I can't run it from my CI/CD since it doesn't have access to the production Postgres (not part of the VPC and shouldn't be), so when should it be done? And how...?

Hello everyone, I’m currently upgrading from prisma1 to prisma2, specifically wanted to move from prisma-binding to nexus. I see that there are 2 prisma plugins to use with nexus:

nexus-plugin-prisma

and

nexus-prisma

. As I understood I’d better use

nexus-plugin-prisma

in production, but wanted to ask how hard will it be to then replace it with the new

nexus-prisma

when it’s production ready ? I mean it seems like

nexus-plugin-prisma

is gonna be deprecated soon, but

nexus-prisma

is not ready, maybe it’s safer to wait until it’s ready and for now use SDL first ? Thank you.

is there any update on this?: https://github.com/prisma/prisma/issues/5040

How is everyone handling slugs in urls? and the logic for when that slug changes? Are you making slugs unique in your model? Consider this scenario: • You create a task, and it has a title. You slugify that and make it unique. So when a user visits /task/{slug} - you can fetch that unique task and get that data. • But when a user changes the title - Do you update the slug? • What happens if that url with the old slug has been linked externally (e.g. on slack, on the web etc) - how do you handle fallbacks? is there best practices to make human friendly urls like this easier to handle?

Hello prisma team! Have a question with respect to prisma connections. How does prisma fetch a connection from connection pool? Context - I am using Postgres as DB and turned logging on. So in the following code snippet, prisma is fetching two connections from pool.

// Node js
// Fetched a connection from pool. --> (1)
const firstQuery = await prisma.model.find();
// Fetched a connection from pool. --> (2)
const secondQuery = await prisma.model.find();

So

1

and

2

statements are logged while calling these queries. Can you explain why second statement is logged? Prisma should use the same connection (connection used for performing first query), right?

Also can we log the current number of used connections?

Hello people Is it possible to make several micro-services communicate with a single instance of the Prisma Client? I'm asking this because the Prisma Client is consuming some considerable memory in each microservice, something that can be prevented if a single instance of the Prisma client is consumed. Instead of: M1 -> Prisma Client 1 M2 -> Prisma Client 2 M3 -> Prisma Client 3 M4 -> Prisma Client 4 I want to get something like: M1 -> Prisma Client 1 M2 -> Prisma Client 1 M3 -> Prisma Client 1 M4 -> Prisma Client 1

Hey guys, just a quick question, is it normal behavior that I'm unable to update a one-to-one relationship on the required side? I'm sure that the update value is correct / unused. When I do, I get this error

The change you are trying to make would violate the required relation...

This may be a complete oversight but thanks in advance for the help! 🙂

Hey guys! I'm new here, learning about Prisma. Is this syntax still valid https://github.com/prisma/prisma/issues/2245#issuecomment-617038028 ?

I'm trying the following and it fails

const test = await this.prisma.proyecto.findMany({
      include: {
        owner: {
          where: {
            nombre: {
              contains: 'asd'
            }
          }
        }
      }
    });

Unknown arg `where` in include.owner.where for type Owner. Did you mean `select`? Available args:
type owner {
}

Prisma1 specially 1.34 with docker latest version on mac M1 no more supports? or wait for something and any ideas for it? Entire project move to Prisma2 I’m not sure also sounds are feel like a so costly. 🤧

Hey guys, is there a way to have composite primary keys defined in Prisma schema?

Is any of the following part of the prisma road map • adding aggregate to the fluent api and batch those requests, so they can be called from the parent model or • allowing aggregate queries to be batched regardless

const User = objectType({
  name: 'User',
  definition(t) {
    <http://t.nonNull.int|t.nonNull.int>('id')
    t.string('name')
    t.nonNull.string('email')
    t.nonNull.list.nonNull.field('posts', {
      type: 'Post',
      resolve: (parent, _, context) => {
        return context.prisma.user
          .findUnique({
            where: { id: parent.id || undefined },
          })
          .aggregate(*some logic)
      },
    })
  },
})