can anyone tell me how i'd dynamically create a transaction spec? i want to do something like:

(defn latest-query
  ([] '[:find (pull ?e [*]) :where [?e :arb/doctype]])
  ([doctype]
   (let [doctype-attr (keyword "doctype" doctype)]
    `[:find (pull ?e [*])
      :where [?e :arb/doctype ~doctype-attr]])))

[:find
 (datemo.routes/pull datemo.routes/?e [clojure.core/*])
 :where
 [datemo.routes/?e :arb/doctype :doctype/note]]

Actually, just figured out one way to do it that's kind of ugly:

(defn latest-query
  ([] '[:find (pull ?e [*]) :where [?e :arb/doctype]])
  ([doctype]
   (let [doctype-attr (keyword "doctype" doctype)]
    [:find '(pull ?e [*])
      :where ['?e ':arb/doctype doctype-attr]])))

Maybe there's a nicer way to do this?

@ezmiller77 what do you mean, 'a transaction spec' ?

@val_waeselynck I just mean the edn for a transaction...

you mean a query, I guess? 🙂

I guess the docs call it "tx-data"

you could always just pass it in as a parameter instead of building a data structure

@ezmiller77 it does look like a query 🙂

Well @val_waeselynck I dont' really care to quibble about symantics.

[:find (pull ?e [*]) :in [$ ?doctype] :where [?e :arb/doctype ?doctype]]

@augustl that's a nice soln!

@ezmiller77 only mentioning it because a query and a transaction are fundamentally very different things

if something takes tx-data it certainly does not expect a query 🙂

@ezmiller77 I don't think it's fair to call it quibbling, you should know the difference just so you can choose between d/q and d/transact

as in, Datomic is not like SQL where you use the same system for reading and writing, they are pretty different

but hey, sorry to have interrupted you with my quibbling. good day!

@ezmiller77 the advantage of parameterizing the query is that datomic will cache the query as well

it means that db and doctype needs to be passed in to latest-query of course

that makes more sense to me at least

I'd also rename doctype-attr to doctype-val 🙂

I would say that the attr here is :arb/doctype

right you are.

Does anybody know if it's possible to backup/restore and rotate log files to an S3-compatible store? I don't see a way of specifying an endpoint URL in the command line tools, and all attempts so far have resulted in the default http://s3.amazonaws.com endpoint appearing in the logged exceptions.

I think the aws build for datomic does this / can do this.

We seem to have a aws-s3-log-bucket-id in aws.properties

afaik it only runs every 24h

@dominicm Yes, I've specified a custom address in the aws-s3-log-bucket-id property but it's ignored. The implementation seems fairly determined that the s3 host name is http://s3.amazonaws.com, whereas I need to specify an alternative host name for a compatible s3 store that we run internally.

ah, I missed s3-compatible.

@kjothen It isn’t currently possible to specify a non-s3 but s3-compatible endpoint for log rotation I’d recommend that you request it as a feature in the customer feedback portal (Suggest Features link in the top nav of my.datomic dashboard)

@marshall I figured this was the case, but thought I'd check nonetheless. AWS is not an option at my workplace just now, but figured I could take advantage of our internal S3 store for backups in particular. I'll raise some feedback - thanks!

@kjothen yep, totally understand the desire/feature - definitely think it’s one to put on the portal

@marshall have any of the portal requests become released features yet?

yep

i believe you can click on ‘released’ or ‘releases’ to see which ones

log in mem db was one

nice