const { data } = await supabase.from<SbConversation>(CONVERSATIONS_VIEW)
      .select("*")
      .match({
        userId: currentUser.id,
      })
.or("lastRead.gt.latestMessage->>createdAt");

I get the response:

invalid input syntax for type timestamp with time zone: \"latestMessage->>createdAt\"

It seems to be selected

createdAt

which is nested inside a

jsonb

object, since it knows it is a

timestamp with time zone

. Perhaps there's something wrong with using

.gt

?

Whenever I go to deploy a function using supabase I get the following error which stalls my CLI (it also breaks my OS and have to do a hard reset).

Error: Get "https://github.com/denoland/deno/releases/latest/download/deno-aarch64-apple-darwin.zip": dial tcp: lookup github.com: i/o timeout

With the release of the v2 JS client and the depreciation of setAuth, I'm currently trying to use the

auth-ui-react

UI component coupled with

supabase-js@RC

and I noticed that the cookie created is only

secure

and not

httpOnly

. Should I be worried about that? Is the only other solution to manually create the cookie server-side with the flag on and send it to the client? Given that setAuth is no longer available, how would I check if user is valid?

Supposing you mean in a client context ie the browser and whether the user themself is logged in, you can check for the supabase-client .auth.session(), if it is not null the user is logged in. In the v2 rc this is now auth.getSession().

When self-hosting your own Supabase instance is it the same UI as offered on

https://app.supabase.com/

?

I'm sure I'm missing something here. I'm using supabase-v2 and loaded Database types using the cli tool. I'm initializing the supabase client like so:

createClient<Database>

But the data returned is of type

undefined[] | undefined

. I'm on supabase-js 2.0.0-rc.10.

I'm relatively new to both Supabase and Svelte Kit. I see there is an Auth Helper for Suapbase with SvelteKit (https://supabase.com/docs/guides/auth/auth-helpers/sveltekit). Maybe I'm just missing something but is this documentation out of date? There are naming conventions and stores in the example that I don't see anywhere else and aren't available such as: -

import { session } from '$app/stores'

-

__layout.svelte.ts

-

hooks.ts

instead of

hooks.server.ts

-

import type { GetSession } from '@sveltejs/kit'

Is it possible to automatically add the user email to another table upon user creation? I need to have a User table with all users information, where this user info's is only manageable by the authenticated user, so the authenticated user can only read and update this specific row.

Hey, I'm implementing a Supabase Function (Deno) using

supabase-js v2

on both client and server. I can create the client with the

ANON_KEY

, receive the user with

client.auth.getUser(bearer-token)

but I fail to set the client context for this user. I need this to have RLS applied to my database calls. How can I create the supabase client on the server on behalf of the user?

supabase.functions.invoke(...)

does not send any cookies, using the authorization header in

createClient(...)

does not work.

For example something like

js
const { data, error } = await supabase
  .rpc('add_task_with_tags', { tags: [{owner_id: 1, title: "tag1"}] })

I'm sending multiple broadcasts per second and every 1 out of 5 broadcast gets skipped randomly , I'm using the paid version. Do I need to configure anything?

I want to add a place for my user settings, and account settings. Is it better practice to create a separate table for this, called something like, "user_settings" and "account_settings" or is it better to just use a JSON column and insert the settings into a "users" public table that I already have.

I am trying to migrate to V2 and I realized the queries were not working at all with V2, so I set up this two test functions: V1:

type group = Database['public']['Tables']['groups']['Row'];

const supabase = createClient(url, anonKey);

const propColumns = 'name, isTrashed:is_trashed';

(async () => {
    await supabase.auth.signIn({email, password});

    const props = 
        await supabase
            .from<group>('groups')
            .select(propColumns);

    console.log(props);

})();

V2:

const supabase = createClient<Database>(url, anonKey);

const propColumns = 'name, isTrashed:is_trashed';

(async () => {
    await supabase.auth.signInWithPassword({email, password});

    const props = 
        await supabase
            .from('groups')
            .select(propColumns);

    console.log(props);
})();

And indeed, the V1 version works as expected, but the V2 one does not return the correct data, it's

undefined

I've taken a look at the logs in the app page and there's this key difference in there: V1:

"sb": [
    {
      "auth_user": "e92a8280-9d20-4dd9-9c0f-40c441810c41"
    }
  ],
  "search": "?select=name%2CisTrashed%3Ais_trashed",
  "url": "https://app url/rest/v1/groups?select=name%2CisTrashed%3Ais_trashed"

V2:

"sb": [],
  "search": "?select=name%2CisTrashed%3Ais_trashed",
  "url": "https://app url/rest/v1/groups?select=name%2CisTrashed%3Ais_trashed"

It seems like the client isn't authenticated with V2, so RLS is doing its job. Is there something I'm missing regarding

signInWithPassword

? The way it is talked about in the release notes sounds like it should be a swap in replacement and a quick look at the source seems to confirm that it does basically the same thing as v1's

signIn

. Would greatly appreciate any pointers

Here is my view:

select l.*, 
"sellerData"."sellerData", "listingLikes"."listingLikes", "lineItems"."lineItems", "listingComments"."listingComments"
from listings l

left join ( 
select id, 
row_to_json(ud.*) as "sellerData" 
from "userData" ud 
group by ud.id 
) 
"sellerData" on "sellerData".id = l."sellerId"

left join ( 
select "userId", 
"listingId", 
array_agg(jsonb_build_object('id', ud.id, 'username', 
ud.username, 'avatar', ud.avatar)) as "listingLikes" from "listingLikes" ll 
join "userData" ud on ud.id = ll."userId" 
group by ll."userId", ll."listingId" 
) 
"listingLikes" on "listingLikes"."listingId" = l.id

left join ( 
select "listingId", 
array_agg(row_to_json(li.*)) as "lineItems" 
from "lineItems" li group by li."listingId" 
) 
"lineItems" on "lineItems"."listingId" = l.id

left join ( 
select "listingId", 
array_agg(row_to_json(cm.*)) as "listingComments" from "listingComments" cm 
group by cm."listingId" 
) 
"listingComments" on "listingComments"."listingId" = l.id

The image attached shows that for every row that should be in

listingLikes

, it returns a duplicate. So if a

listing

has 5 likes, 5 duplicate rows will be in my view.

I did the following:

sql
ALTER USER postgres SET search_path TO "$user", public, auth;

don't even remember exactly where and what for, just that it was based on a comment suggestion to "fix" something. Now I'd like to reset it to default but I'm not sure what the default is supposed to be. I'm totally new to Postgres, have some experience with other SQL DBs 🤓 Found this

sql
ALTER ROLE postgres SET search_path TO "\$user",public,extensions;

in Would that do? Or, if the value is actually correct better to do

sql
ALTER USER postgres SET search_path TO "\$user",public,extensions;

Reading , my last example might be the right one tho still not completely sure 😬

I have a view which is combination of 3 tables, I call this view with anonymous token. I added one of the tables a RLS policy that simply returns

FALSE

. I was expecting request I made from a page should fail but to my suprize it didn't . What am I missing here?

So it seems like I should be able to order by nested values on the supabase js sdk according to this: https://github.com/supabase/supabase/discussions/229

typescript
q = this.sb.supabase.from('tags')
  .select('*, pid!inner(*, author!inner(*))', { count: 'exact' })
  .eq('name', tag)
  .order('pid.created_at', { ascending: true })
  .range(from, to)

tags - pid (fk to posts.id) - ... posts - created_at - ... I'm trying to sort by the posts created at field. I get a 400 error. However, when I comment out the

.order()

line, everything works fine. J

This sections seem to missing, let know what am doing wrong.

when converting a downloaded blob from storage to a URL, i get the error

TypeError [ERR_INVALID_ARG_TYPE]: The "obj" argument must be an instance of Blob. Received an instance of Blob

Hello, I wanted to know how I could run the following SQL query if I'm using the Supabase JavaScript SDK:

sql
SELECT * FROM project WHERE email = 'example@email.com';

Hi! I have a table called

FriendsList

and i am trying to build a query to fetch the friends of a specific user,

FriendsList(sender, recipient)

so it is a bit tricky since the current user can be either a sender or a receiver and its giving me a headache since i am not familiar with supabase queries. Any tips would be greatly appreciated 🙏

Is there a way to reference environment variables from a postgresql function for use in an http request? I have dev, staging, and prod environments and I don't want to hardcode the urls into the function. Do postgresql functions have access to supabase secrets? Particularly the host url of the supabase project and the api keys? I'd like to use this to call and edge function from the db when a record is inserted.

I'm not looking to modify the user, but want to create one if one doesn't already exist.

Hi folks. I've been going back and forth on this and I'm at a loss. I've done my best to setup a local dev environment with the CLI so that I can test and develop locally before pushing to the remote db. But without parity between the local dev environment and the online production environment I keep running into issues. The big one is functions and webhooks. I've tried working within my limitations and using what is available in the local dev env to create triggers and functions to replicate webhooks but the setup is brittle. I've also tried setting up my functions and webhooks through the online/production dashboard and then run

supabase db remote commit

on a fresh local environment but the migration doesn't include the supabase_functions schema or any of the associated tables and functions. So what would be a reasonable workflow to separate testing from production? I don't want to make changes to the production db while I'm testing. Would I create two projects online - one for production and another for testing? And if so, how can I push changes between the two? Or, do I use the CLI and create custom functions to mimic the supabase functionality?

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;

Following the docs in setting up my first edge function, and ran into

{"message":"no Route matched with those values"}

when trying to hit the endpoint. Can't find any info about it online

Hi, where do I get the REALTIME_URL?

Hello I'm trying to automatically update the user table with the avatar_url of whatever auth service the user logs in with. Does anyone have a working solution or tips on how to do this?

Hey all, wondering if anyone has any suggestions on how to call a supabase edge function at 'second' intervals? Im currently using the cron extension to trigger intervals but it currently only supports the standard cron minute granularity.