SStupid me... example is on Wrangler 2.15, I used 3. Will revert back to 2.15 as used in the sample. Therefor yes...
JSo I was able to bind my DB to my worker script. Is there a good page with example queries? Basically I want to run this from my worker script: SELECT * FROM DB_BIND WHERE name LIKE %city% LIMIT 10; then output the results in json, or false if it doesn't find anything.
JSadly chat GPT doesn't understand D1 yet 
JI tried this but it doesn't work:
try {
const result = await CITIES_ALL.prepare(
return new Response(JSON.stringify(result), {
status: 200,
headers: {...headers, 'Content-Type': 'application/json'}
});
} catch (error) {
return new Response(JSON.stringify({ error: 'Error' }), {
status: 500,
headers: {...headers, 'Content-Type': 'application/json'}
});
}
try {
const result = await CITIES_ALL.prepare(
select * from cities_all where name like ? limit 10
).bind('%' + sanitizedCity + '%').all()return new Response(JSON.stringify(result), {
status: 200,
headers: {...headers, 'Content-Type': 'application/json'}
});
} catch (error) {
return new Response(JSON.stringify({ error: 'Error' }), {
status: 500,
headers: {...headers, 'Content-Type': 'application/json'}
});
}
JThe error I get is "ReferenceError: CITIES_ALL is not defined", so I perhaps I am not initializing it properly? I got this message when I published it, so I think I must be part of the the way there: Your worker has access to the following bindings:
- D1 Databases:
- CITIES_ALL: cities_all xxxx
Dyou're trying to access it as
CITIES_ALL, which is the syntax for service workers - the new module worker syntax makes bindings available under env that you get in the handler, so you should be using env.CITIES_ALLJI tried env.CITIES_ALL as well, and it got the same error:
'ReferenceError: env is not defined'
'ReferenceError: env is not defined'
Dcan you show your worker code?
Dyou're using service worker syntax, which doesn't work with D1 - here's how to migrate: https://developers.cloudflare.com/workers/learning/migrating-to-module-workers/#migrating-to-module-workers
This guide will show you how to migrate your Workers from the Service Worker syntax to the new Module Worker format.
JAh, ok thanks for that. I'll take a look!
JHmmm, maybe I am misunderstanding something and the database is only available locally?
JI probably should wait until D1 is out of alpha anyway before I continue tinkering with it. Thank you for your help!
Bhow do i bind d1 to a worker, i need a basic backend and db for my school project :p
MOrdering requires a larger read, unless your index is ordered.
MYou’d likely want a multi-column index - on both (id, created_at) - https://developers.cloudflare.com/d1/learning/using-indexes/#multi-column-indexes
Indexes enable D1 to improve query performance over the indexed columns for common (popular) queries by reducing the amount of data (number of rows) …
Mhttps://use-the-index-luke.com/sql/sorting-grouping/order-by-asc-desc-nulls-last is a good read. LIMIT is only limits what is returned. Not read.
Indexes that optimize ORDER BY must consider the ASC/DESC and NULLS FIRST/LAST modifiers.
Jis there going to be a way to move a database from like planetscale to d1? And move from d1 to something else if d1 for some reason doesnt suite my needs anymore. Just wanna make sure of no vendor lock
MRegression. Tracking internally. Will aim to fix this week.
MMigrating between databases - PlanetScale is MySQL based and D1 is SQLite based - is not always trivial. Query languages aren’t the same, nor are types. For simple use-cases it can be easy.
MDownloadable as? A BLOB is binary data - at best we can give you an
octet-stream but you’d need to determine what the data actually is/what application can handle it.
Mi currently have a use case where I need to interact with ~20 DO's concurrently, and I think I'm getting some poor performance due to the maximum of 6 concurrent connections. Might migrating to D1 solve this issue?
IIt depends. D1 doesn't lift the limit of 6 concurrent connections, but if you can move all of the work on the ~20 DOs to a single D1 DB then that might work for you. It entirely depends on what your data/schema looks like though - some data isn't a great fit for SQL.
MThanks Isaac, I need to map an input string to a UUID globally with strong consistency. Once that mapping is created, it's placed in KV and read from there until the end of time, i.e. that input string no longer will pass through DO
However I need to do this mapping in chunks of 20-50 with strong consistency, not sure what type of solution I am looking for here
However I need to do this mapping in chunks of 20-50 with strong consistency, not sure what type of solution I am looking for here
ID1 is strongly consistent as it stands right now. However when replicas are made available sometime in the future those replicas will likely be eventually consistent.
II'm not sure whether there will be an option to force a read from the primary - I would hope so but I don't know for sure.
IAre you using the new backend or the old backend?
I(if you don't know or created your DB before 3 days ago you are using the old backend)
IInteresting. I'm not sure whether that's expected - someone else here would need to confirm.
