Cloudflare Developers•3y ago

but if D1 gets charged for creating a database with no rows then there goes my theory

Ccaptain but if D1 gets charged for creating a database with no rows then there goes my t...

Hello, I’m Allie!•1/28/23, 10:33 PM

We also have no info for how D1 is billed atm, so we can’t comment there

Unsmart•1/28/23, 10:33 PM

DOs are only charged when you have an active request being processed. The duration charge when all you do is storage is basically nothing and I normally dont include in my cost estimates tbh (other than maybe some rounding to the next $10/$5 lol)

Unsmart•1/28/23, 10:35 PM

Fwiw D1 is built on top of DOs so for any cost estimates I would almost assume it probably costs the same or more

(but thats just my opinion and I have no idea what it actually will be)

UUnsmart DOs are only charged when you have an active request being processed. The durati...

captainOP•1/28/23, 10:35 PM

got it. Well I have yet to play with one, but for now i was toying with the D1 idea...thought i was gonna have it done on monday...but then had to convert my whole giant project to modules from a server worker

Original message was deleted

captainOP•1/28/23, 10:36 PM

i learned that this week lol

captainOP•1/28/23, 10:36 PM

but now my project is much faster and takes way less time to build

captainOP•1/28/23, 10:36 PM

so its good

Noel•1/29/23, 10:01 AM

just came here to ask this ^

CCarlson Orozco Hello everyone, does anyone know if there's an update on this? https://community...

Max (@rozenmd)•1/29/23, 2:54 PM

it's in the backlog, I've been nudging to get it done

Walshy•1/29/23, 3:36 PM

@yuimaru Removed your message since you posted your Turnstile key but as for why that env is there, that's how the D1 shim needs the env to be. It should be available at

env.DB

env.DB

still and work great
If you're using a _worker.js then you'll need to bundle it

--bundle

--bundle

Original message was deleted

James•1/29/23, 3:37 PM

What's up?

kev-ac•1/29/23, 9:18 PM

Maybe API bug: I have one D1 database which is at capacity (~104MB, 1 Table). When I try to open it in the dashboard the "tables" tab thinks some seconds and then nearly always fails with

API Request Failed: GET /api/v4/accounts/{ACCOUNT_ID}/d1/database/{DATABASE_ID}/table (500)

API Request Failed: GET /api/v4/accounts/{ACCOUNT_ID}/d1/database/{DATABASE_ID}/table (500)

.

API Response:
{
"result": null,
"result_info": null,
"success": false,
"errors": [
{
"code": 7501,
"message": "Internal error"
}
],
"messages": []
}

James•1/29/23, 9:26 PM

I'm seeing similar errors on a couple of my test D1 databases too, even just hitting

/d1/database/dc00d2ec-2c06-460d-a6e4-01cfe42f93e4

/d1/database/dc00d2ec-2c06-460d-a6e4-01cfe42f93e4

on the dashboard

Erisa•1/30/23, 12:49 PM

afaik that API is undocumented and could break/change at any time, and also im pretty sure it would be slower latency-wise in a lot of cases and has rate limiting applied

Erisa•1/30/23, 12:50 PM

if its just an occasional query thats not latency-sensitive youll be fine, but if youre querying on every request for example i would strongly recommend using the worker binding

Erisa•1/30/23, 1:11 PM

very very useful for one-offs, but yep definitely dont rely on it for production traffic

Erisa•1/30/23, 1:33 PM

none that im aware of

Erisa•1/30/23, 1:33 PM

its mostly useful to workers, which dont really support c#

Erisa•1/30/23, 1:34 PM

you could, but keep in mind the above caveats
generally i would recommend instead using a worker as a proxy for d1, rather than using the api for anything more than one-offs

Max (@rozenmd)•1/30/23, 7:49 PM

also new in wrangler 2.9.0:

wrangler d1 execute --jsonwrangler d1 execute --json
wrangler d1 list --jsonwrangler d1 list --json

meow•1/31/23, 9:22 AM

I presume that .run().run() is for inserts and deletes, but I notice that it has a resultsresults property on the return type which is never actually used. When I try to use the RETURNRETURN statement in SQL I never get results, unlike running the same insert query with .all().all().

44x4notfound i'm getting 800ms+ response times from d1 today. wondering if there is something...

Max (@rozenmd)•2/1/23, 7:13 AM

I wouldn't monitor performance during the alpha

Max (@rozenmd)•2/1/23, 7:18 AM

enabling replication will likely change the performance characteristics quite a bit, and we're still evaluating the approach we've used to run sqlite within core, so... soontm

meow•2/1/23, 9:37 AM

Is there a way to parameterize the response type of a .batch().batch() when the statements have more than one return type? For example .batch<T1, T2>(stmt_1, stmt_2).batch<T1, T2>(stmt_1, stmt_2)? Or does .batch().batch() expect queries with homogeneous return types?