when you apply a migration for an experimental backend db, (for now) you need to pass `--experimenta

when you apply a migration for an experimental backend db, (for now) you need to pass

--experimental-backend

--experimental-backend

--experimental-backend

--experimental-backend

MMax (@rozenmd)when you apply a migration for an experimental backend db, (for now) you need to...

yeehow•6/1/23, 7:22 PM

ahhh ok thank you

Original message was deleted

Leon338•6/1/23, 9:21 PM

I'm not entirely sure but logically it has to scan every CompanyName or atleast the first digit, because even the last row could be this CompanyName. Could be fixed by setting a Limit at the End for example to find only the first one. But yes an Index sorted by CompanyName should help alot. Basically would change from O(n) to O(log(n))

Leon338•6/1/23, 9:26 PM

And keep in Mind the Reads are Size based. If your CompanyNames has an average length of 8 chars it would take 500 scans for one Read Unit. But maybe the Where statement is optimized so that it doesnt even read the entire CompanyName

Original message was deleted

Leon338•6/1/23, 9:34 PM

Nevermind I think im wrong with the Index. The WHERE LIKE statement can be different and also search if something contains anything. I don't even think a Index would help there and it's always O(n)

Original message was deleted

Leon338•6/1/23, 10:10 PM

Not if you have a % at the start and at the end, thats always an entire Table Scan

Leon338•6/1/23, 10:16 PM

But you could try to use Contains instead of Like that should be faster and is more optimized I guess

Leon338•6/1/23, 10:20 PM

And if the Company Name contains multiple Words you can look into a Fulltext Index, but I mean this type of Queries will always scan alot of rows

UUnsmart As I have already said in the past the pricing difference entirely depends. If y...

Matt Silverlock•6/1/23, 11:51 PM

You could fully scan a 1GB table 100,000 times before you reach the included 25 billion reads. Read units might add up fast but we give you billions and charge 1/10th of a penny per million.

Unsmart•6/1/23, 11:52 PM

That is true, I dont know what the cost for each would be for a normal usage as an actual database was mostly pointing out that KV is meant for larger storage blobs where if possible to store on D1 wouldnt be suitable cost wise

Unsmart•6/1/23, 11:53 PM

Probably shouldve made that more clear to be fair

MMatt Silverlock You could fully scan a 1GB table 100,000 times before you reach the included 25 ...

Matt Silverlock•6/1/23, 11:53 PM

D1 should always be cheaper for SQL/queryable type workloads. If you’re treating it like a big KV store with > 1MB rows then yeah, it’s not a good fit. That’s OK - that’s what KV is built for.

UUnsmart That is true, I dont know what the cost for each would be for a normal usage as ...

Matt Silverlock•6/1/23, 11:53 PM

Entirely agree.

MMatt Silverlock D1 should always be cheaper for SQL/queryable type workloads. If you’re treating...

Leon338•6/2/23, 1:17 AM

My thought was, if I use D1 to store Session Tokens (sorted) it would be 100x cheaper then KV. Even tho KV is basically the perfect fit for it.

LeftSpace•6/2/23, 11:43 AM

does d1 works on miniflare (local version)

LLeftSpace does d1 works on miniflare (local version)

Max (@rozenmd)OP•6/2/23, 11:47 AM

yep

LeftSpace•6/2/23, 11:56 AM

couldn't find on documentation. do i have to configure with wrangler.toml?

LLeftSpace couldn't find on documentation. do i have to configure with wrangler.toml?

Max (@rozenmd)OP•6/2/23, 11:57 AM

are you on wrangler 2 or 3?

LeftSpace•6/2/23, 11:58 AM

3.0.1

LeftSpace•6/2/23, 11:59 AM

A local session does not have access to your production data by default. To access your production (remote) database

LeftSpace•6/2/23, 11:59 AM

i see

Max (@rozenmd)OP•6/2/23, 11:59 AM

if you run

wrangler dev

wrangler dev

, it'll be accessing a local miniflare d1 instance

LeftSpace•6/2/23, 11:59 AM

so if i pass --remote

LeftSpace•6/2/23, 11:59 AM

it'll use remote db

Max (@rozenmd)OP•6/2/23, 11:59 AM

should be, yeah

LeftSpace•6/2/23, 12:00 PM

thanks >.<

LeftSpace•6/2/23, 12:01 PM

Any changes you make when running in --remote mode cannot be undone.

LeftSpace•6/2/23, 12:02 PM

so this will effect my already running remote code?

LeftSpace•6/2/23, 12:02 PM

i dont want that if thats the case.

LLeftSpace Any changes you make when running in --remote mode cannot be undone.

Hello, I’m Allie!•6/2/23, 12:04 PM

No, it’s more that if you accidentally break your database by using some buggy code locally, then you can’t just rollback when you stop the dev session. It doesn’t affect your deployed code, and once the D1 backup restore thing, then you can roll it back manually

LeftSpace•6/2/23, 12:05 PM

ah thats fair. yeah i'll just read data wont create any table n stuff Im good then.

Mmarshallswain 🎁 I made a FeathersJS integration for Cloudflare D1, this week, running Kysely...

Matt Silverlock•6/2/23, 12:08 PM

This is awesome. Want to PR it here? https://developers.cloudflare.com/d1/platform/community-projects/

LLeftSpace couldn't find on documentation. do i have to configure with wrangler.toml?

Matt Silverlock•6/2/23, 12:08 PM

https://developers.cloudflare.com/d1/learning/local-development/

MMatt Silverlock https://developers.cloudflare.com/d1/learning/local-development/

LeftSpace•6/2/23, 12:12 PM

thanks i made it working already!

Original message was deleted

Matt Silverlock•6/2/23, 6:04 PM

LIKELIKE and GLOBGLOB can use indexes - see The LIKE OptimizationThe LIKE Optimization

https://www.sqlite.org/optoverview.html

Ggotfredsen ok what is the current state regarding primary key in D1 .. is simple auto incre...

Matt Silverlock•6/2/23, 6:05 PM

https://developers.cloudflare.com/d1/learning/using-indexes/#create-an-index

see the note there. If you don't specify a primary key there is an implicit ROWID INTEGER PRIMARY KEYROWID INTEGER PRIMARY KEY that is used.

Ggotfredsen ok what is the current state regarding primary key in D1 .. is simple auto incre...

Matt Silverlock•6/2/23, 6:06 PM

If you're doing lookups by some form of UUID, then you should create an index on that column so that your WHERE order_id = ?1WHERE order_id = ?1 queries benefit from the index.

Matt Silverlock•6/2/23, 6:12 PM

The implicit ROWIDROWID is unique, but (per my comments above):

If you're doing lookups by some form of UUID, then you should create an index on that column so that your WHERE order_id = ?1 queries benefit from the index.

Take a look at the docs in full - you can also define a UNIQUEUNIQUE index to prevent duplicate UUIDs (which should be near impossible, but an application bug on your side could re-use the same UUID) - https://developers.cloudflare.com/d1/learning/using-indexes/#when-is-an-index-useful

when you apply a migration for an experimental backend db, (for now) you need to pass `--experimenta

Similar Threads