and is there anyway I can restore database from a local backup?
and is there anyway I can restore database from a local backup?
Myou can expect a performance improvement, here's what we noticed:
Myou'll need to dump the local backup to sql first: https://developers.cloudflare.com/d1/learning/importing-data/#converting-sqlite-database-files
TThanks.
Mthat was just execution
Mwe have yet to find a solution to beating the speed of light
Myeah that ones kinda pesky
Nis it possible to get the query output as csv in wrangler?
Mjust JSON atm
Mif you know you can convert JSON input to CSV output in your terminal, though
Nthanks, looks like a good tool that i can use in the future
Nhowever this time i just ended up using a js snippet from so to convert the json in the browser console 
Mthe fix for this should land in 3-4 hours
Ccould someone clarify me some things about D1? Is it based on sqlite, and, if yes, where can I find some list with supported features? Does it have support for foreign keys? can it be used with something like drizzle orm? thank you.
BCheck out the T4 stack: https://app.t4stack.com/ I am working with it now and using Drizzle with it. I struggle with what is available vs. not too, but FKs are supported. There isn't yet support for transactions.
Tamagui, Solito, Expo & Next.js
Cthank you for the recommendation and the input. I kinda wish the D1 documentation was more explicit. I've found in the meantime on another page that D1 is based on SQLite, but I believe this is something that should be mentioned in the D1 docs. Anyway, I assume this is because D1 is still in alpha.
CHaving FKs at least is good, means we won't end up with something like planetscale again
HDrizzle specifically supports d1, you can see how to set it up in their documentation (https://orm.drizzle.team/docs/installation-and-db-connection/sqlite/d1). Foreign keys are also supported (as sqlite have support for foreign keys). I agree with you planetscales lack of support for FKs was such a pain
NHey everyone, when a D1 is created and the replicas distributed through the network. I wanna know if that D1 is always up and running or if it only runs when a request has been made to it?
Cif I were to answer by intuition, given that D1 is based on SQLite it doesn't actually "run" anywhere – it's just a file on each node of the network. "Replication" of D1 is basically copying the database file across nodes, just as the static assets of a website would be. For the database to "run", I assume that a Worker reads the database file on request, if necessary, and works with that.
In essence, you could say that D1 "runs" only when a request is made. But it really never runs, as it is not a server. Pricing also reflects that (https://developers.cloudflare.com/d1/platform/pricing/): notice that the metrics by which usage is priced are reads, writes, and storage (which are characteristic to file usage), but not something like CPU time (which is characteristic to a server).
In essence, you could say that D1 "runs" only when a request is made. But it really never runs, as it is not a server. Pricing also reflects that (https://developers.cloudflare.com/d1/platform/pricing/): notice that the metrics by which usage is priced are reads, writes, and storage (which are characteristic to file usage), but not something like CPU time (which is characteristic to a server).
CIf somebody spots a mistake please correct me.
NPerfect that makes sense and thank you for all the other tid bits of information.
HD1 is not currently replicated, and iirc they hibernate after around 30 seconds of inactivity
NSo, a typical PostgreSQL requires a server to be up and running at all times to serve records. But since D1 is built on SQLite - from what I understand, SQLite is serverless and doesn't require a server - it's just a file like CiabbataConda said.
just learning all this
just learning all this
NI intended to use PostgreSQL for the project I started building and now I am wondering if I can use D1 instead
HSort of. Standard SQLite is just a file plus some code to read it. But how D1 works is that it is some code that Cloudflare runs for you, on top of that file. So yes, that service may need to be cold-booted occasionally if it hasn’t been called in a while, but as Skye pointed out, bootups occur in the order of a few milliseconds, so it shouldn’t be anything you need to factor into your plans
Cwhere do you get this info from? is there some documentation on the internals which I am incapable of finding?
Ci'd like to read more about this technology but god I can't find anything on it
HThe Cloudflare Blog
Today, we’re excited to announce D1, Cloudflare’s first SQL database, designed for Cloudflare Workers
NI'm gonna google more about the diff. between PostgreSQL and SQLite to understand more
HThe Cloudflare Blog
We've been heads down iterating on D1, and we've just shipped a major new version that's substantially faster, more reliable, and introduces Time Travel: the ability to restore a D1 database to any point in time.
Cof course they're blog posts...
Cthank you
NI'm not sure I even understand the whole complexity involved in a business like DoorDash but if someone were to build an application with DoorDash's business model, I don't know how many but say they needed 10-15-20 tables with 7-10 columns... would D1 be able to handle that?
Cthis is built with D1: https://northwind.d1sql.com/. here is the database: https://github.com/jpwhite3/northwind-SQLite3. you can find there a graph of its schema.
MTime Travel is now available in the latest
wranglerTime Travel is D1’s approach to backups and point-in-time-recovery, and allows you to restore a database to any minute within the last 30 days.
Cthis is rather a question of whether SQLite can handle scale. see https://www.sqlite.org/whentouse.html, https://news.ycombinator.com/item?id=33975635.
MThat is not a lot of tables. Schema complexity isn't really a problem. Making your queries efficient - D1 or not - is what is important.
NThat schema png is exactly what I had in mind. I'll check those links out, thanks.
NI'll keep that in mind
N"To use SpatiaLite with SQLite, you need to install the SpatiaLite library and load it as an extension to SQLite."
Does D1 currently have the ability to add a library as an extension?
Does D1 currently have the ability to add a library as an extension?
BAre databases created via the dashboard on the experimental backend? Or is that just an option via the CLI? And how can you tell which backend you're using?
CIt used to be an option when creating to enable experimental, but it looks like it was changed so that they're all experimental now. A really easy way to tell is if the dashboard has "Backups" or "Time Travel" option inside of the database. If it's time travel, it's running experimental.
Also if it has the metrics tab, only experimental has that.
Also if it has the metrics tab, only experimental has that.
BThanks!
TIs there any way to share/access a D1 Database from multiple CF accounts, from different workers or pages from different accounts?
CThere is not. Resources like KV, R2, D1 are all scoped by account so you can only share between workers of the same account
TConfirmation of what I was afraid... Thank you for the answer
MD1 Dashboard seems to have trouble with BLOB columns, it appears to take the column's representation as a Uint8Array and concatenate all of the decimal values therein, resulting in this to represent a column containing :
(please forgive my horrendous screenshot skills)
(please forgive my horrendous screenshot skills)
