Hi, my app has regular logged in users, but I'd need some users to be admins so that they can access the backoffice. From the UI i'd check if they are admin and let them see the backoffice, from supabase I will also create policies that will only allow queries in the backoffice to users with role admin. How would you change the role? I was not able to find anything about this. Thanks

Is there somewhere I can see documentation of what each oauth provider returns as metadata? I'm trying to work out what data I'll be able to save for each user when they sign up but I can't find any documentation about the different metadata each provider returns.

so im doing google o auth, which works fine, but unfortunately im not able to dispatch them to react context like this ( google auth just redirects me back to the home page )

Hi everyone, I'm currently implementing a node js api for my react front and want to use the supabase rest api with axios. But I cannot find how to get the count of a table ? If I'm using thenpm package every call is made via Websockets and it cannot work on my desktop. thanks for your help

Had some issues with the API this morning, looked in the dashboard and can't access most options. Only one project has been affected by this. Error below: Error: [500] failed to get pg.tables: could not open file "global/pg_filenode.map": Input/output error

I am trying to upload a file to a storage bucket using the JS library. I think I have set the right policies for insert and select. Why would i be getting this error?

For compliance reasons, we need to self-host. We're an AWS-based shop and would prefer to use AWS Aurora PostgreSQL for our backend instead of a docker or EC2 based Postgres database. The problem is, we don't have access to the postgres user and all the Supabase init scripts assume you are running as that user. We were able to work around that by making some small tweaks to the scripts and preparing an admin-style account on our own ahead of time, but when we try to start up the realtime container, the eval Realtime.Release.migrate portion of the command fails.

Here's a scenario. Say you have a project with a few functions (e.g. 30). Common code has been refactored in a utility.js/.ts file so that it can be accessed from some of these functions. At some point, a bug is fixed in this file which is used in say, 20 of these files. Now... in Firebase, it's up to the developer to trace back the calls and figure out the top functions/endpoints to deploy. This is time-consuming, non-productive and error-prone. How about Supabase? Is there a way to deploy everything so no function gets left behind? I see in the docs that each function must be uploaded individually? If so, it sounds pretty limiting. What is your workflow?

Hey, I am currently wondering if there is any kind of best practices for indices when setting up rls for storage? Status Quo: Since the folder path is concatenated in the "name" property of the storage.objects table I do not see a way to setup an rls and indice where the DB just look at the first or second part of the name path in a performance efficient way. **Example**: folder path: //file_name We have an RLS Setup that checks if the user has access to the team_id Now I would like to setup an indice so that a user can access the files of his team very efficient, as currently we are performing very week index scan, which are actually full table scans. Is there any kind of example for RLS and indice setup available or any tip to setup a RLS on folder level with the corresponding indices?

hey I installed the supabase v2 RC in my project which uses pnpm, but there's no binary found. other binaries seem to appear just fine

Error querying the database: db error: FATAL: remaining connection slots are reserved for non-replication superuser connections

We're only getting this when we switch from Railway to Supabase. See pic of logs. Any idea why?

sql
BEGIN
  IF new.bucket_id != 'team-gallery' THEN
    RETURN null;
  END IF;
  INSERT INTO public.gallery
  (object_id, user_id, team_id)
  VALUES
  (new.id, new.owner, new.path_tokens[0]::uuid);

  RETURN new;
END;

While inserting the team_id into the gallery table from the new.path_tokens[0] team_id is null. The object that I am listening for an insert from is in a folder with the uuid of the team.

I'm currently the founder of a global app with thousands of users. We want to limit latency for our global users but also want to migrate to Supabase!

Gets my free-tier app suspended if no requests are made to it, or if no admin logs in to the console?

Does Supabase Graphql support aggregations? J

Hi there Would it be possible to setup a flow where a user is invited via zapier? I have no idea how it would work, but I'd like to be able to have it set up so when [insert thing] happens off site, a user with the corresponding mail gets invited. Is this too out there or is it actually possible?

Hi! I am trying to set the credentials with

supabase.auth.setAuth(token)

, and the subscriptions in

realtime.subscription

table vanish after a few seconds. In logs I found this:

"Received an invalid access token from client: {:error, :unknown}" "Unknown connection authorization error: :error"

. I use the flutter SDK and the jwt is signed using jwt secret and the payload contains all the required claims like for example

{ "supabase_uid": "1ed9dcf5-e383-4e41-820f-58df106c3e5c", "role": "authenticated", "aud": "authenticated", "app_metadata": {}, "user_metadata": {}, "iss": "my_supabase", "iat": 1663875937, "exp": 1663880737 }

I test my code on a self-hosted supabase and it works as expected. Can anyone please tell me what can be done in this case?

Is it possible to move a project from one organization to another organization in same account?

select *
from profiles 
INNER JOIN images ON profiles.id = images.profile_id
INNER JOIN follows on profiles.id = follows.following_id
where follower_id = 12;

This works:

let { data: images } = await supabase
        .from('images')
        .select('*, profiles!inner(*)')
        .range(0, 25)
        .eq('profiles.id', 10)
        .order('created_at', { ascending: false });

But can't figure out the syntax to add follows.following_id. It is a FK to profiles but it's a composite key with followers_id. For some reason can't get a working version of this SQL query into Javascript.

I'm running supabase-auth-helpers-react on next.js and we have a login page that emails the user a magic link. Upon authorization when the link is clicked, the user should be redirected to the '/profile' page using: supabase.auth.signIn( { email: data.email }, { redirectTo:

${process.env.NEXT_PUBLIC_HOST_DOMAIN}/profile

} ) When the profile page is fetching the user data client side by hitting an api endpoint within the app, the redirect works accordingly. const { user, error } = useUser() useEffect(() => { async function loadData() { fetch(

/api/users/${user?.id}

) .then(res => res.json()) .then(result => setInfo(result)) } // Only run query once user is logged in. if (user) { try { loadData() console.log(user) } catch (error) { console.log(error) } } }, [user]) When fetching the data using getServerSideProps, the redirect no longer works and i get sent to the home route of the app. The user is confirmed to be authorized and signed in when navigating back to the /profile page. export const getServerSideProps = withPageAuth({ async getServerSideProps(ctx) { const prisma = new PrismaClient() // Access the user object const { user, accessToken } = await getUser(ctx) const prismaUser = await prisma.user.findUnique({ where: { id: user?.id, }, }) return { props: { user, prismaUser, }, } }, }) Has anyone else experienced this redirect bug when fetching data server side rather than client side? Any help would be greatly appreciated!

I'm trying to make a View with my hearts count: hearts - pid (fk to posts) - ... posts - id -...

sql
CREATE OR REPLACE VIEW posts_hearts_count AS
SELECT hearts.pid, posts.*, COUNT(hearts.pid) AS heartsCount
FROM posts
JOIN hearts ON posts.id = hearts.pid
GROUP BY hearts.pid

But I get this error:

Failed to run sql query: column "posts.created_at" must appear in the GROUP BY clause or be used in an aggregate function

I basically want my entire posts data along with the

heartsCount

field for the total likes of that post. Relearning SQL is interesting... J

I am thinking of using Superbase in my next project. I would have to import a large data set of data to the database. Am I able to easily add a property and value to every object in the database ? the property name and value would be the same for every object.

You should grant the role that is trying to access from the JavaScript client to the custom schema. Grant …. To (role) On (custom schema) (Syntax sth like above. I didn’t verify it)

I came across the

supa_audit

package which looks sufficient, but the issue is we manage authentication outside of supabase so we have an issue of not knowing what user made a specific change to a given record/row... Whats the best approach to solve this?

How could I relate information such as a caption, team_id (uuid of the team that the image belongs to), location, and other stuff to the image in storage. I am looking for a way that won't leave a image with no data or data with no image. (AKA I don't want to upload the image, get the id from the image, then create the record all client side. Is there a way to do this with RPC or a edge function? (One invoke/call from browser to create record and upload image?)

Why is the OAuth screen page using the Supabase project URL instead of the App name i specified in google console's oauth consent screen settings? Is it because it is the Authorized redirect URIs for my OAuth 2.0 client id on google console settings?

Curious if realtime works for database views or only database tables?

Is it possible to invoked edge functions from web client? Currently I am adding a new row to a table to invoke the generation of a new Twilio Access Token for a user, is this the easiest way to do so currently in your opinion?

Hey everyone, I am having major struggles with the query performance of the storage.objects table and cannot get my head around it. Any help is highly appreciated. We are providing a multi-tenant solution where different entities can have files attached. To easily query them on different levels, we enforce the following directory structure:

organisation-id/entity-id/myfile.jpg

In some cases, we also have one more level, e.g. for conversation with messages, we want to attach files to a message but also query all files of a conversation:

organisation-id/conversation-id/message-id/myfile.jpg

. To query files for an entity, we use the following query (simplified):

sql
select id
from storage.objects
where bucket_id = 'message_files'
and uuid_from_path_level_2(name) = '7daa4a53-1233-432d-bf11-a9dee8140027'::uuid;

For a better performance, we setup a bunch of indices, and when executed as service_role (bypassing rls), the indices are properly picked up ([Visualised on explain.dalibo](https://explain.dalibo.com/plan/4247g6e82726c139)).

Hey supabase frens, I saw that in the latest comment that sign in with ethereum / other web3 auth ideas are depriroritzed right now for good reasons: httpsgithub.comsupabasegotruepull282. I saw

moralis

and

clerk

offering some third party soultions as well as someone suggesting to just use

wagmi

or

rainbowkit

and link a new web3 table to the original

users

table via a foreign key. Anyone has tried any of this? What was your experience? Would love to implement an open source solution but haven't found any good tutorial or codebases. Would love to hear your experience or any links to REPOS where one implemented this! Thanks 🙏