Fork me on GitHub
#datomic
<
2017-10-28
>
lmergen09:10:31

so i'm relatively new to Datomic, and trying to figure out what makes the most sense. if i have an entity with a unique identity (a uuid) that spans multiple domains, would it be better to use two different datomic identities (created using d/tempid), or is it safe to just use one single identity that spans multiple domain barriers ? for the sake of example, let's consider the case where i have one customer id that's used by both the accounting department and infrastructure department. so i have :accounting/name, :accounting/customer-id and :infrastructure/foobar values, that are all uniquely identified by a customer id. how would you design this ?

lmergen09:10:32

also, i definitely need to be able to pull the :accounting/* information when doing :infrastructure queries, but not vice versa. i guess this influences design.

the-kenny11:10:26

lmergen: My guess is that a single entity with attributes from both domains is the correctâ„¢ thing to use.

the-kenny11:10:52

But you have to make sure there aren't any conflicts possible, now and in the future

the-kenny11:10:36

As for pulling stuff: As long as there's a reference between two entities you can just use the entity api and walk the tree in any direction

chris_johnson15:10:04

Just to confirm what I expect to be true (or maybe learn something that makes my morning much better hehe) - if I want to run a transactor in AWS and I want a custom logback.xml for something other than dumping logs to S3, is it the case that I have to roll my own CF template rather than going the bin/create-cf-template route and using a Cognitect-provided AMI?