Cloudflare Developers•2y ago

I'm thinking about it, maybe not if workers work like lambda functions

RRubi yes, but are you still gonna deploy in vps? you gonna use the http api version

KimitriOP•4/29/24, 8:35 AM

The problem is, I use an Orm (typeOrm) and I don't know if I can connect D1 with typeorm

KKimitri The problem is, I use an Orm (typeOrm) and I don't know if I can connect D1 with...

Rubi•4/29/24, 8:38 AM

yeah, right now, using the http api, i dont think so

KKimitri I'm thinking about it, maybe not if workers work like lambda functions

Rubi•4/29/24, 8:52 AM

worker node support is more limited than lambda function

Ccopper not really, just take something like hono which runs on workers (and pages?) and...

KimitriOP•4/29/24, 8:55 AM

Yeah, I'm new with cf workers and d1, but I really interesting in it because looks so much faster than VPS in one region.

RRubi worker node support is more limited than lambda function

KimitriOP•4/29/24, 8:57 AM

for checkout, files compress, and ecommerce is workers recommended?

KKimitri for checkout, files compress, and ecommerce is workers recommended?

Rubi•4/29/24, 8:59 AM

if my orientation is deploying on pages/worker, I would choose any library that support edge environment, if not, I'll figure it out outside pages/worker

KimitriOP•4/29/24, 9:02 AM

thank you for clarify, I'll keep read the documentation to learn how d1 and workers work to do the migration without some erros for currents endpoints.

Rubi•4/29/24, 9:05 AM

lol, that ^ is enough to get started

KimitriOP•4/29/24, 9:06 AM

wow good idea, I'll make it and when it's done I come back to show the results

Marcel Overdijk•4/29/24, 9:34 AM

I'm developing an application for Cloudflare that will use D1 as database.
It's semi-static data that will not change frequenly. Bi-weekly I must basically re-import the whole database (~250.000 records).

What's the best option, during that process just drop the tables and re-import.
Or create a new database, import the data, update the binding and redeploy/restart? the Cloudflare Pages/workers app?

MMarcel Overdijk I'm developing an application for Cloudflare that will use D1 as database. It's ...

Rubi•4/29/24, 9:37 AM

how much downtime your app could tolerate?

Rubi•4/29/24, 9:38 AM

this might be dirt in the code, but I might do: create new table name, import the data, put the table name to env, use retry deployment or not at all

Marcel Overdijk•4/29/24, 11:42 AM

A bit of downtime is no problem at the moment, but if I could avoid it that would be a plus.

MMarcel Overdijk A bit of downtime is no problem at the moment, but if I could avoid it that woul...

Rubi•4/29/24, 1:37 PM

yeah, I would use the new table and query using join with dynamic table name

SELECT *
FROM (
    SELECT latest_table_name
    FROM table_source
) AS dynamic_table
JOIN table_name ON dynamic_table.value = table_name.table_name;

SELECT *
FROM (
    SELECT latest_table_name
    FROM table_source
) AS dynamic_table
JOIN table_name ON dynamic_table.value = table_name.table_name;

Oom Is there a way i can use D1 for this? Like download the D1 db through workers, a...

Rubi•4/29/24, 1:42 PM

1. download the D1 db through worker, idk, but wrangler able to do that by

npx wrangler d1 export dbname --remote --output db.sql

npx wrangler d1 export dbname --remote --output db.sql

so there is a way
2. someone can upload and use it, the D1 REST API should do the job

Kotamigo•4/29/24, 3:29 PM

Hello everyone! I'm currently integrating SvelteKit with Prisma and D1 for a web application, utilizing pages and workers. I've set up authentication with Google and have a corresponding auth handle in hooks.server.ts. However, I'm encountering an issue post-authentication:

After the Google auth redirect to http://localhost:5173/auth/callback/google with the necessary query params for code and scope, I try to access my database configuration from the platform environment (event?.platform?.env?.DB) within the hook, but it returns as undefined. Does anyone have insights on why this might be happening or how to ensure the DB variable is properly set and accessible within this context?

For all other requests, the platform attribute is present on the event object. I have been able to keep local development moving using

getPlatformProxy

getPlatformProxy

imported from wrangler, but this doesn't work in production because wrangler depends on node libraries that aren't available in production. I need the database in this context to create user records (they are created by an auth.js prisma adapter)

Related code: https://pastebin.com/ynzs24Jh

Thanks in advance for any help or pointers!

Pastebin

export async function getPrismaClient(event?: RequestEvent): Promise

Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.

Ccopper just read your comparisons with ps, have the same experience (ps is ultra fast a...

Super Morris•4/29/24, 5:02 PM

hello cooper, our benchmarks are from last week. however my dev did some magic with planet scale to keep the connection open and stuff (i dont really know) to achieve this results

feras•4/30/24, 3:31 AM

Hello everyone, is there a way to expose sqlite functions like sqlite3_update_hook?

kian•4/30/24, 3:39 AM

Not currently - it looks like it was requested last year and at the time there was no plan to implement a feature like that (https://community.cloudflare.com/t/watching-for-changes/476125/3)

feras•4/30/24, 3:44 AM

what about triggers in sqlite, are they supported?

kian•4/30/24, 4:12 AM

Yes

feras•4/30/24, 4:17 AM

thank you kian, is there any documentation on this?

kian•4/30/24, 4:29 AM

Nothing D1 specific, just https://www.sqlite.org/lang_createtrigger.html

feras•4/30/24, 5:06 AM

thank you for your answers, is there access to sqlite3_create_function?

kian•4/30/24, 5:29 AM

Nope

kian•4/30/24, 5:29 AM

You only have access to what's on https://developers.cloudflare.com/d1/build-with-d1/d1-client-api/

feras•4/30/24, 6:05 AM

it would've been awesome to have functions or update hook to listen to changes

Fferas it would've been awesome to have functions or update hook to listen to changes

Max (@rozenmd)•4/30/24, 6:59 AM

we're just getting started with D1 fwiw

MMax (@rozenmd)we're just getting started with D1 fwiw

feras•4/30/24, 7:25 AM

great job so far, hope to see full features soon

Ccopper does d1 sleep?

Hello, I’m Allie!•4/30/24, 12:56 PM

They do, yes, though you aren't billed for whether the D1 instance is asleep or not

Hello, I’m Allie!•4/30/24, 12:56 PM

You only pay for storage/ops

Ccopper ok and do you know the time d1 needs to wake up before it can respond to a reque...

Isaac McFadyen•4/30/24, 4:34 PM

It's fairly low - on the order of 10s to 100s of ms. I believe the more you have stored the longer it will be, but still fairly low in any case.

James•4/30/24, 5:22 PM

Quick question, and I know alpha databases should be avoided, but doesn't "deprecation" not mean "completely disabling functionality"?

From the recent email:

On July 1, 2024, we will deprecate D1 alpha databases. Upon deprecation, alpha databases will reject live SQL queries

I don't personally have any alpha DBs that I use in production anyway, but describing this as a "deprecation" but then completely disabling all queries to it seems to counter eachother a little. This was the first real direct comms I've received about D1 alpha DBs being formally deprecated. cc @Vy

James•4/30/24, 5:43 PM

To be clear, I'm not against the deprecation of the old alpha DBs, but this seems to be the first formal comms I've received about it. I totally understand preventing new ones being created, and actively discouraging their usage even, but a comms window of ~2 months until all queries just start getting rejected feels a little hasty. What if someone was on vacation for the next month or two, and hasn't been actively following discord or places where this deprecation may have been discussed? cc @Matt

I'm thinking about it, maybe not if workers work like lambda functions

Similar Threads

Similar Threads

Similar Threads