It doesn't break compatibility with the old API, you can still use the old API just fine. It's just

Zzegevlier It doesn't break compatibility with the old API, you can still use the old API j...

L

LordSilver•4/3/24, 9:16 PM

So you can use both at the same time? Just one will make use of the deployed replicas while the other won't?

LLordSilver So you can use both at the same time? Just one will make use of the deployed rep...

Z

zegevlierOP•4/3/24, 9:18 PM

I can't confirm whether you'll be able to use them simultaneously, but D1 is a GA product, so I would not expect any of the current APIs to change in a backwards-incompatible way.

U

Unsmart•4/3/24, 9:20 PM

If theres any changes it will be guarded by a compat date/flag

A

albyzyx•4/3/24, 9:55 PM

Hey team, What is the best implementation to achieve realtime using d1? some sort of CDC or access to the WAL?

Zzegevlier It doesn't break compatibility with the old API, you can still use the old API j...

M

Matt Silverlock•4/3/24, 9:58 PM

Exactly this.

You can use the existing

env.DB.prepare()

env.DB.prepare()

env.DB.prepare()

env.DB.prepare() and other APIs, or you create a session via

let session = env.DB.withSession(commitToken)

let session = env.DB.withSession(commitToken)

let session = env.DB.withSession(commitToken)

let session = env.DB.withSession(commitToken) and then call

session.prepare()

session.prepare()

session.prepare()

session.prepare(),

session.batch()

session.batch()

session.batch()

session.batch(), etc, etc.

We aren't breaking the APIs. Users that never adopt the session API will have D1 as it works today, with the same consistency model they've built their app against.

Developers who want to use replicas can adopt the session API.

F

forslunds•4/3/24, 10:02 PM

Do we really need replicas if we go 1db/user

if we dont have heavy writes ofc

U

Unsmart•4/3/24, 10:05 PM

replicas only helps with reads not writes

J

johtso•4/3/24, 10:57 PM

are there any tools that let you explore your d1 database with a gui similar to drizzle studio?

J

johtso•4/4/24, 12:13 AM

going to have a play with d1-manager, curious to hear if anyone's had any experience with it or has any other suggestions

Jjohtso are there any tools that let you explore your d1 database with a gui similar to ...

M

mulch•4/4/24, 3:46 AM

from this post it looks like it's possible to use Drizzle Studio with D1 (if you're using Drizzle): https://kevinkipp.com/blog/going-full-stack-on-astro-with-cloudflare-d1-and-drizzle/

Kevin Kipp

Going full-stack on Astro with Cloudflare D1 and Drizzle

A step-by-step guide to adding a back-end to your Astro project using Cloudflare D1 and Drizzle ORM

M

mulch•4/4/24, 3:47 AM

Any ETA on worker-level support for the per-user DB approach?

Mmulch Any ETA on worker-level support for the per-user DB approach?

M

Max (@rozenmd)•4/4/24, 5:31 AM

we don't do ETAs

Original message was deleted

F

forslunds•4/4/24, 5:32 AM

True your case need it, good luck

i only have isolated users that dont/cant see other data so it should always be closest to the user

Fforslunds Do we really need replicas if we go 1db/user 🤔 if we dont have heavy writes ofc

M

Max (@rozenmd)•4/4/24, 6:06 AM

read replicas are also critical above the 500-1k RPS mark, otherwise the individual DO starts queuing/overloading

MMax (@rozenmd)read replicas are also critical above the 500-1k RPS mark, otherwise the individ...

F

forslunds•4/4/24, 6:19 AM

Aah right its an DO, valid case! Thanks for heads up. Cant imagine that load per user for my case but better safe than sorry

(especially with that “price”!)

CChaika Looks like Oceania/Australia DOs are online! https://where.durableobjects.live/c...

M

Max (@rozenmd)•4/4/24, 8:05 AM

:soontm:

LLiudmila cold start time with D1 appears to be 2sec, my DB is <2MB. Is this expected? Do ...

M

Matt Silverlock•4/4/24, 12:51 PM

This is definitely not expected - we have metrics for cold starts. Are you seeing this repeatedly, or just a one-off?

L

Lucho•4/4/24, 2:42 PM

hey, im a beginner with D1 and i was hoping to find some help
I created a model for a DB with prisma and migrated it SQLite. How can i "connect" this DB to Cloudflare D1?

LLucho hey, im a beginner with D1 and i was hoping to find some help I created a model ...

M

Matt Silverlock•4/4/24, 3:01 PM

I'd recommend stepping through this guide from beginning to end! https://developers.cloudflare.com/d1/tutorials/d1-and-prisma-orm/

Cloudflare Docs

Query D1 using Prisma ORM · Cloudflare D1 docs

Prisma ORM is a next-generation JavaScript and TypeScript ORM that unlocks a new level of developer experience when working with databases thanks to …

L

Lucho•4/4/24, 4:17 PM

thank you!!

L

Lucho•4/4/24, 6:05 PM

I was following the guide until this error appeared

L

Lucho•4/4/24, 6:06 PM

does someone know how to resolve it?

LLucho I was following the guide until this error appeared

M

Max (@rozenmd)•4/4/24, 6:09 PM

whats the contents of that SQL file?

L

Lucho•4/4/24, 6:10 PM

just this

LLucho just this

M

Max (@rozenmd)•4/4/24, 6:11 PM

out of curiousity, does it work if you remove the comments?

-- CreateTable

-- CreateTable

and

-- CreateIndex

-- CreateIndex

L

Lucho•4/4/24, 6:12 PM

i'll try

L

Lucho•4/4/24, 6:12 PM

nope

L

Lucho•4/4/24, 6:12 PM

same error

M

Max (@rozenmd)•4/4/24, 6:17 PM

and the contents of that file is now just this?

CREATE TABLE "User" ( 
    "id" INTEGER NOT NULL PRIMARY KEY AUTOINCREMENT, 
    "email" TEXT NOT NULL, 
    "name" TEXT 
);

CREATE UNIQUE INDEX "User_email_key" ON "User"("email");

CREATE TABLE "User" ( 
    "id" INTEGER NOT NULL PRIMARY KEY AUTOINCREMENT, 
    "email" TEXT NOT NULL, 
    "name" TEXT 
);

CREATE UNIQUE INDEX "User_email_key" ON "User"("email");

CREATE TABLE "User" ( 
    "id" INTEGER NOT NULL PRIMARY KEY AUTOINCREMENT, 
    "email" TEXT NOT NULL, 
    "name" TEXT 
);

CREATE UNIQUE INDEX "User_email_key" ON "User"("email");

CREATE TABLE "User" ( 
    "id" INTEGER NOT NULL PRIMARY KEY AUTOINCREMENT, 
    "email" TEXT NOT NULL, 
    "name" TEXT 
);

CREATE UNIQUE INDEX "User_email_key" ON "User"("email");

LLucho just this

T

tsion•4/4/24, 6:33 PM

the loose emoji on line 10 is probably the problem?

Ttsion the loose emoji on line 10 is probably the problem?

L

Lucho•4/4/24, 6:34 PM

it isnt an emoji, it's from copilot

L

Lucho•4/4/24, 6:35 PM

it goes away if i tap aywhere else

T

tsion•4/4/24, 6:35 PM

i see

MMax (@rozenmd)and the contents of that file is now just this? ``` CREATE TABLE "User" ( "...

L

Lucho•4/4/24, 6:38 PM

thats right

I

Isaac McFadyen•4/4/24, 6:57 PM

Check your line endings?

I

Isaac McFadyen•4/4/24, 6:58 PM

The "near ??" error makes it seem like there are some unprintable characters in there somewhere, maybe line endings or empty spaces?

M

Matt Silverlock•4/4/24, 6:58 PM

Runs perfectly for me as copied from @Max (@rozenmd)'s code block:

~/repos/d1 (main) ✗
➜  pbpaste > user_repro_20240404.sql
~/repos/d1 (main) ✗
➜  wrangler d1 execute db-enam --file=user_repro_20240404.sql

🌀 Mapping SQL input into an array of statements
🌀 Parsing 2 statements
🌀 Executing on db-enam (c89db32e-83f4-4e62-8cd7-7c8f97659029):
🚣 Executed 2 commands in 0.4695ms

~/repos/d1 (main) ✗
➜  pbpaste > user_repro_20240404.sql
~/repos/d1 (main) ✗
➜  wrangler d1 execute db-enam --file=user_repro_20240404.sql

🌀 Mapping SQL input into an array of statements
🌀 Parsing 2 statements
🌀 Executing on db-enam (c89db32e-83f4-4e62-8cd7-7c8f97659029):
🚣 Executed 2 commands in 0.4695ms

~/repos/d1 (main) ✗
➜  pbpaste > user_repro_20240404.sql
~/repos/d1 (main) ✗
➜  wrangler d1 execute db-enam --file=user_repro_20240404.sql

🌀 Mapping SQL input into an array of statements
🌀 Parsing 2 statements
🌀 Executing on db-enam (c89db32e-83f4-4e62-8cd7-7c8f97659029):
🚣 Executed 2 commands in 0.4695ms

~/repos/d1 (main) ✗
➜  pbpaste > user_repro_20240404.sql
~/repos/d1 (main) ✗
➜  wrangler d1 execute db-enam --file=user_repro_20240404.sql

🌀 Mapping SQL input into an array of statements
🌀 Parsing 2 statements
🌀 Executing on db-enam (c89db32e-83f4-4e62-8cd7-7c8f97659029):
🚣 Executed 2 commands in 0.4695ms

I

Isaac McFadyen•4/4/24, 7:00 PM

If you're on a Windows system try changing from CRLF to just LF?

M

Matt Silverlock•4/4/24, 7:02 PM

Something is definitely off with your editor here, @Lucho.

The migration (just copy pasted from the docs):

K

kian•4/4/24, 7:09 PM

at offset 0 is weird

K

kian•4/4/24, 7:10 PM

it's pulled from

sqlite3_error_offset

sqlite3_error_offset

which indicates it's the first character of that given input, but since Wrangler splits the statements you don't really know which one it is

I

Isaac McFadyen•4/4/24, 7:24 PM

Just tried CRLF vs LF and both work so that's not the issue.

K

kian•4/4/24, 7:28 PM

Usually if it's LF/CRLF you get a "you can't execute multiple statements" error or something like that

K

kian•4/4/24, 7:29 PM

A prepared SQL statement must contain only one statement. even

MMatt Silverlock Something is definitely off with your editor here, @Lucho. The migration (just ...

L

Lucho•4/4/24, 7:39 PM

oh okay

L

Lucho•4/4/24, 7:39 PM

it might be

L

Lucho•4/4/24, 7:41 PM

i'll try doing the entire guide again from scratch

A

adjlsncm•4/4/24, 8:11 PM

is d1 sqlite or sql? is sqlite worse?

A

adjlsncm•4/4/24, 8:15 PM

it seems like sqlite is worse for anything potentially large scale from what im reading

It doesn't break compatibility with the old API, you can still use the old API just fine. It's just

Similar Threads