We do not have a backup feature in datomic cloud. Is this purely for a disaster recovery scenario?

@U1QJACBUM Yes. If something happens to the data, I would like to be able to restore them

@U1QJACBUM Can you recommend something?

Hi @U028H6X0KRS as I mentioned we do not have a backup feature in Datomic Cloud like our backup/restore feature in Pro. We have on-going work in this area to provide a feature, but I cannot provide a timeline for that work. It is important for me to point out that the stratified durability of Datomic Cloud perists data onto DynamoDB, S3, EFS and any data missing from any given layer can be recovered from deeper within the stack. We talk about this a bit in the docs here https://docs.datomic.com/cloud/whatis/architecture.html#stratified-durability. Of course that does not satisfy many DR plans (i.e. a region goes down). You can write your own backup tool by reading from the current system and writing to another system. You can also utilize https://docs.datomic.com/cloud/dev-local.html#import-cloud to move your cloud DB locally via dev-local, but understand these solutions do not satisfy every DR requirement nor are they a great fit for extremely large DBs.

Hi @U1QJACBUM. Can there be problems if I use the import-cloudbetween the clouds (copying from production, for example, to staging)?

@U028H6X0KRS import-cloud takes a datomic system and imports it to local disk for dev local. You cannot then sync it up to another system. The functionality goes in one direction.

Only by using the transaction entity

so not really

A datom in datomic is [E A V TX], where TX is the entity id of the transaction. You can assert facts on the transaction entity; if you are ok with a little semantic hand-waving, you can have those transaction assertions be understood as actually asserting something about facts in the transaction

but this is the closest you can get; otherwise you need to reify the datom as an entity

thank you favla, i don't know how easy it will be, or how to match the TX in queries, for example how to model this in datomic ? [[:question1 :got-answer :answer1] :points 100]

:find ?points :where [:question1 :got-answer :answer1 ?tx] [?tx :points ?points]

I doubt this is what you want though--this is a hack, because your assertion is on the transaction entity; which means you must transact facts on boundaries which correspond to how you want to assert more things about them--in the degenerate case, a fact per transaction

you also can’t pull (via pull expression) through the transaction entity of a datom

this mechanism is really meant for metadata about the transaction itself

ok thank you for helping me favila, i am just thinking rdf database like jena or datomic

datomic is RDF-inspired, but it’s not an RDF database nor does it claim to be

e.g., it’s closed world, has no reification, doesn’t implement any RDF specifications

yes, the obvious is to like choose datomic because clojure, but on RDF so many specifications, people behind tools etc and i dont have the knowledge to compare and decide for me

i will see : )

people generally use RDF for interop reasons, not because it’s fun to build standalone applications in it

anyway, if you need rich consumption or production of RDF data, or reasoners over it, probably datomic is going to add significant friction to that. If you just want to build a clojure application and need an application database, datomic is probably a better bet because it isn’t distracted by the very large surface area of RDF specifications and stuff

thank you, i will probably see those in practice : )