Yep, cloud query cannot destructure in :find

(Actually client api query)

If datomic cloud doesn't have the collection find spec, what is the recommended way to return a collection instead? Massage it on the programming language level?

yes

that’s all it was doing anyway.

[?a …] => (map peek results)

[?a ?b] => (first results)

?a . => (ffirst results)

Just the general unix advice? Don’t daemonize: manage the process with an init system like systemd

would also recommend a systemd unit

got it. yep that's exactly the advice I was looking for. Also I thought systemd was a daemonization process?? Maybe I'm just confused on terminology. I meant "robust way to run the thing in the background better than a screen session"

are you just running a one-off transactor in a terminal? screen, new terminal tab, or even bash job-control is fine

are you running in production as a service? use a process manager

by “don’t daemonize” I mean don’t do the double fork/join “daemon” mode of older style inits

where the transactor gets detached from its parent process

^ +100

let systemd sort it out

systemd supports daemonization but it's not the preferred default service type

We uploaded dev-local to a private s3 repo and bring it in with tools-deps s3 support. I think lein has a s3 wagon to do the same.

A query can take in multiple dbs.

There was a limitation, but seems to be solved now, can you confirm it @U083D6HK9?

Good news indeed!

Depends on what limitation you're referring to.

@U083D6HK9 I read in the current #datomic channel and here also: https://forum.datomic.com/t/filtering-and-multi-database-queries/481 that cross join is not allowed when one is using Datomic Cloud.

(Just following this conversation in case there has been an update, as we’ve been handling joins in client code)

Great @U4Z2TFMJ9, I’m reading the Datomic Cloud documentation and seems to be the case. Sometimes is a little bit painful to revisit all docs to know if something changed. Maybe the Release Notes is the best place to be aware of all changes.

We have code that queries Datomic like the last query in that blog post.

Yeah, thanks @U28A9C90Q. I think @U083D6HK9 rightly pointed out that query can take in multiple dbs, however it appears in cloud those dbs must be the same.

(d/q {:query '[:find (count ?a)
               :in $1
               :where
               [$1 _ :db/ident ?a]]
      :args [db-1]})
=> [[661]]
(d/q {:query '[:find (count ?a)
               :in $2
               :where
               [$2 _ :db/ident ?a]]
      :args [db-2]})
=> [[68]]
(d/q {:query '[:find (count ?a)
               :in $1 $2
               :where
               [$1 _ :db/ident ?a]
               [$2 _ :db/ident ?a]]
      :args [db-1 db-2]})
Execution error (ExceptionInfo) at datomic.client.api.async/ares (async.clj:58).
Query db arg does not match connection

Exactly @U35LPCHNH

Oh, right - I misread the q. Our queries only ever cross the same connection.

> Sometimes is a little bit painful to revisit all docs @U28A9C90Q They post releases to https://forum.datomic.com/c/datomic-gen/announcements/6. You can follow it for updates.