I was using D1 for the first time in a while and came across a D1 error that was confusing: ```types

I was using D1 for the first time in a while and came across a D1 error that was confusing:

env.PREPARE("SELECT ...").bind(...).run()

env.PREPARE("SELECT ...").bind(...).run()

env.PREPARE("SELECT ...").bind(...).run()

env.PREPARE("SELECT ...").bind(...).run()

This doesn't work because

run()

run()

run()

run() apparently isn't allowed for

SELECT

SELECT

SELECT

SELECT (but I forgot I needed to use

all()

all()

all()

all() or

first()

first()

first()

first()).
Unfortunately, all that's thrown is

D1_ERROR

D1_ERROR

D1_ERROR

D1_ERROR (with nothing in

e.cause

e.cause

e.cause

e.cause.)

Are there some docs that help explain the JS API? Or maybe some docstrings could be added to the types to help explain when to use each method?

JacobOP•4/10/23, 2:21 PM

That's true, deff could have been logging it wrong. Thanks!

JJacob I was using D1 for the first time in a while and came across a D1 error that was...

DaniFoldi•4/10/23, 2:31 PM

Types could also be converted into generics for D1PreparedStatementD1PreparedStatement to indicate what statement it was more precisely, something like https://tinyurl.com/d1-generic-types-ts-playground

TS Playground - An online editor for exploring TypeScript and JavaS...

The Playground lets you write TypeScript or JavaScript online in a safe and sharable way.

DDaniFoldi Types could also be converted into generics for `D1PreparedStatement` to indicat...

Matt Silverlock•4/10/23, 3:20 PM

cc @geelen

Original message was deleted

DaniFoldi•4/10/23, 7:33 PM

yeah there are definitely edge cases, but I think there are <10k possible combinations of statement types, so it might be doable - the playground is just an example for me to check if it's possible without regex types literals, so I only did some very basic tests

hanpolo•4/11/23, 5:33 AM

I'm experimenting with running D1 migrations off of Pages CI/CD using wrangler apply with an API token for D1. When I ran the command it applied the first script and then failed on the second script with Internal Error 7501. Does anyone know if there are limitations to doing this: e.g. it can only run one script at a time etc.

hanpolo•4/11/23, 5:36 AM

FYI: I ran the same scripts later from the terminal from my local development mac and everything ran smoothly with changes made to my D1 instance on Cloudflare: so I know that the scripts do not have bugs.

Cc becker This is very common when using wrangler to apply migrations - it will fail partw...

hanpolo•4/12/23, 11:29 PM

Yep. Thanks for the response. Of course this means that we cannot push out code that works with migrations. In such a case I will opt to run migrations manually to make sure the database is updated and then push code to GitHub to update the pages code. Real shame though as it breaks the possibility of truly automated CI/CD

hanpolo•4/13/23, 3:34 AM

I really do not believe in having dozens of if then switch cases strewn all over the code looking at schema. I believe that's a recipe for unmaintainable code. However, I'll go for the practice of pushing out schema changes that work with the old code, then pushing out code that works with the new schema and then deleting the old schema. However, even here, i'd rather do it by pushing schema changes to GitHub and then having CI/CD implement the changes because anytime you have a human do manual commands on a live site you introduce a massive risk factor of mistakes and failure. So IMHO having Wrangler randomly fail and thinking that this is normal does not reflect well on the perception of reliability and production readiness. However, having said all this, I have faith that given D1 is only Alpha, these problems will be resolved hopefully when D1 goes beta. Having Wrangler reliably work with CI/CD is essential to a developer system that can be relied on and is ready for prime time.

TigersWay•4/13/23, 10:31 PM

Could we dream?

https://github.com/nalgeon/sqlean

GitHub

GitHub - nalgeon/sqlean: The ultimate set of SQLite extensions

The ultimate set of SQLite extensions. Contribute to nalgeon/sqlean development by creating an account on GitHub.

TigersWay•4/13/23, 11:19 PM

In the last few hours, a simple insertinsert which took an - somehow stable - average of 25ms for the first 5000+ records (6 days), seems to take now a - more floating - average of 160ms...
While I understand we are not yet in the performance stage of D1, I'm wondering if something is happening!

TTigersWay In the last few hours, a simple `insert` which took an - somehow stable - averag...

Matt Silverlock•4/14/23, 12:23 AM

Alpha performance slows down as your DB gets bigger.

TTigersWay In the last few hours, a simple `insert` which took an - somehow stable - averag...

Matt Silverlock•4/14/23, 12:24 AM

This will NOT be the case during the beta. Trust us. Alpha is really an alpha.

TigersWay•4/14/23, 1:40 AM

@elithrar I was hoping so

.... and I'll go on waiting!

kw•4/14/23, 5:53 PM

Can I sign up for larger database in alpha stage?

kw•4/14/23, 5:53 PM

200MB is not enough for my use cases.

Original message was deleted

kw•4/14/23, 6:17 PM

2GB. I think I read it somewhere for reaching out for larger db.

kw•4/14/23, 7:22 PM

I see. Is there any timeline for beta version?

Kkw I see. Is there any timeline for beta version?

Hello, I’m Allie!•4/14/23, 7:33 PM

hanpolo•4/14/23, 10:01 PM

Question on db.batch(). Can you run prepared statements on two different tables? If so, can you get a result back that has rows from each table? Ideally, I'm looking at inserting rows in two different tables where one table has an Foreign Key relationship with the other. Is this possible with db.batch()?

I was using D1 for the first time in a while and came across a D1 error that was confusing: ```types

Similar Threads