I’m getting an error Caused by: com.mysql.cj.jdbc.exceptions.PacketTooBigException: Packet for query is too large (16,642,439 > 4,194,304). You can change this value on the server by setting the 'max_allowed_packet' variable. trying to restore to an Aurora (mysql) database in AWS

hi all, im struggling to understand the notion of distinct values and aggregates

i have a :db.type/float attribute

if I have, say, three entities with [1.2, 2.0, 4.0], running (sum) on that produces the expected 7.2

but if I have [1.2, 1.2, 1.2], (sum) produces “1.2” and count produces “1"

I dont get what I am doing wrong

any help appreciated

@ghaskins Datomic generally group bindings on value - treating them as a set of values :with can help with controlling the grouping http://docs.datomic.com/query.html#with

@ghaskins always remember that Datomic is set based. This has bitten me tonnes of times. You sort of have to drum that into your head. So you if you get identical tuples, and don't use :with then they will appear as one tuple.

with datomic rules, how would i pass the db as an arg to rules, and then to a function in a rule? like so:

[(chat-about-subject $ ?subject ?chat)
     [?subject :email ?email]
     [(= (some-fn $ ?chat)
         ?email)]]

i keep getting “unable to find symbol $” in this context

or even just pass the db to a function

Is there a library that can generate specs from a schema tx?

Hey all, is there a way to free up resources associated with a datomic connection so that my process doesn't hang on exit? I have a lein task that looks just like this:

(ns slow-lein-because-datomic.core
  (:require [datomic.api :as d])
  (:gen-class))
(defn -main []
  (println "about to connect")
  (d/create-database "datomic:)
  (d/connect "datomic:)
  (println "connected"))

@kbaribeau The hang is caused by the clojure agent pool, which things in datomic use. (This is a general clojure problem, not datomic specific). See docs for datomic.api/shutdown, which may help you.

also clojure's shutdown-agents function

Aha, thanks!

say we a unary ref, its spec would have be an s/keys but it doesn’t feel right to statically define which keys are required and which are optional because we often have different “schemas” on read or write