Cloudflare Developers•3y ago

hey, is it possible to enforce d1/sql table/row policies?

CConstantin hey, is it possible to enforce d1/sql table/row policies?

Matt Silverlock•11/25/23, 9:38 PM

No, D1 doesn’t support row-level security. Not something we plan in the near term.

Bulbasaur•11/25/23, 9:47 PM

is D1 globally read replicated at all yet? Or still just single-region?

Bulbasaur•11/25/23, 9:48 PM

And is that theoretically planned to be a 1.0 feature or... later?

BBulbasaur And is that theoretically planned to be a 1.0 feature or... later?

Matt Silverlock•11/25/23, 9:49 PM

d1-database

Bulbasaur•11/25/23, 9:51 PM

Gotcha cool, I didn't see any mentions in the dos, it might be helpful to have a list of the "still to be implemented" features (unless I missed it)

AAlexander I believe I was never able to access the D1 Dashboard, I get all ways the Worker...

Rob M.•11/25/23, 11:21 PM

Looks like you need to create a worker/page first.

divby0•11/26/23, 1:01 PM

Changing from wrangler v2 to v3, I experience the inability to insert as many values to the db in a single statement as before.

Previously I could add even 50 entries at once, now I can add 16. When I add more than 16 (17 eg) I get no error but the await never resolves.

Is this expected behaviour? I could batch/chunk the insert if needed I just don't know if this is caused by something else maybe

divby0•11/26/23, 1:11 PM

From the limits documentation, I get this:

Maximum SQL statement length    100,000 bytes (100 KB)
Maximum bound parameters per query    100
Maximum arguments per SQL function    32

Maximum SQL statement length    100,000 bytes (100 KB)
Maximum bound parameters per query    100
Maximum arguments per SQL function    32

I have 6 columns, so 6*16 is 96 parameters. This would mathematically perfectly align, but I can't believe that we can never insert more than a few elements at a time for such a big company like cloudflare - also it was possible in v2?

kian•11/26/23, 1:27 PM

If you hit one of those limits, it would result in an error - not a hanging promise.

divby0•11/26/23, 1:27 PM

could this be that drizzle swallows the error maybe?

Kkian If you hit one of those limits, it would result in an error - not a hanging prom...

divby0•11/26/23, 1:32 PM

Also would it be expected for inserting 17 items with 6 columns to not work?

Ddivby0 Changing from wrangler v2 to v3, I experience the inability to insert as many va...

Matt Silverlock•11/26/23, 1:45 PM

Are you running locally?

MMatt Silverlock Are you running locally?

divby0•11/26/23, 1:52 PM

Yes correct

divby0•11/26/23, 2:03 PM

Wrangler v2 was able to run locally as well as on the actual remote worker with more than 16 elements . Easily 50.

I also didn't upgrade the database to the new one yet (or migrate) because of the expression depth problem. I was just grading up wrangler v3 to test out th eexpression depth which now seems to be sufficient

Ccardyet Does anyone know about read-replication. I'm in Lima, Peru and I have pocketbase...

Cyb3r-Jak3•11/26/23, 3:00 PM

Planned for GA but doesn't exist yet

Epailes•11/26/23, 4:28 PM

On the topic of read replication:
Is it known if replicas will be automatically distributed to major cloudflare locales? Or will it be a similar behaviour to cloudflare cache (how frequently is the resource requested? How large is it? How busy is the target locale?) and potentially be in the smalled edge locales?
I'm planning on switching to D1 once it reaches GA, and curious whether it's worth planning to share D1 instances per tenant as that means each instance is more likely to be cached? Or whether that won't have any major impact? Each of my tenants don't use up much data, <10MB.

Ccardyet Does anyone know about read-replication. I'm in Lima, Peru and I have pocketbase...

johnswarbrick•11/26/23, 8:22 PM

114ms is nice. Getting around 250ms for simple SELECT queries on an almost empty DB from here in the UK. Looking forward to the read-replica replication to see whether D1 is a really a viable, super-scaleable alternative to other cloud database providers.

schwartzmj•11/26/23, 10:00 PM

Anyone know if it's possible to run wrangler d1 migrations applywrangler d1 migrations apply during the Cloudflare Pages build step? And if so how to specify the database name at the end of the command dynamically based on a

.env

.env

or some binding

SSuchACharles Running Nextjs 14 edge pages to display data from D1, it works, however, after a...

Matt Silverlock•11/27/23, 12:03 AM

wrangler dev

wrangler dev

is local by default, so you aren’t seeing the local development changes reflected in your “global” D1 database

Jjohnswarbrick 114ms is nice. Getting around 250ms for simple SELECT queries on an almost empty...

Rob M.•11/27/23, 12:06 AM

250ms from the API or worker?

kian•11/27/23, 12:18 AM

What is your Worker's code?

kian•11/27/23, 12:23 AM

Yep!

kian•11/27/23, 12:25 AM

Have you re-generated the Drizzle schema since adding these new columns?

kian•11/27/23, 12:25 AM

Drizzle doesn't use SELECT *SELECT *, it lists all columns from your schema, so it may not be selecting new columns if it hasn't been generated again.

kian•11/27/23, 12:40 AM

No problem, glad it's sorted.

Santosh•11/27/23, 4:40 AM

can someone provide an example on this topic: How do I insert and then append a json object (not a scalar,) but a proper json object using json_insert at the end of an existing json array in D1? My objective is to store data like this in one text column: history: [{object1 with key and values}, {object2 with keys and values}]

Elodere•11/27/23, 10:12 AM

Does D1 support UUIDv7 or should I go with ULID for now?

EElodere Does D1 support UUIDv7 or should I go with ULID for now?

Rob M.•11/27/23, 10:29 AM

Why? UUIDv7 is still in a WIP state itself. Is it a game changer?

Elodere•11/27/23, 10:33 AM

UUIDv7 from what I've glanced of the spec is supposed to be a lot better than v6 (generally recommended to use v7 whenever possible basically)

RRob M.Why? UUIDv7 is still in a WIP state itself. Is it a game changer?

Elodere•11/27/23, 10:33 AM

I don't know enough to claim that it's a game changer, but it's definitely a significant improvement over previous versions

Elodere•11/27/23, 10:34 AM

hence why some people are looking at UUIDv7 as a possible replacement for using ULID

EElodere UUIDv7 from what I've glanced of the spec is supposed to be a lot better than v6...

Elodere•11/27/23, 10:44 AM

Also something I've realized is that the dashboard only seems to allow for string, int, blob, and real datatypes

But because D1 runs on SQL, I assume it should be fine to use datatypes like UUID right?

Elodere•11/27/23, 10:46 AM

Thank you I didn't realize it was SQLite

Elodere•11/27/23, 10:47 AM

Yeah then I was right in just using string as the type

hey, is it possible to enforce d1/sql table/row policies?

Similar Threads

Similar Threads

Similar Threads