datomic odds and ends
Switch branches/tags
Nothing to show
Failed to load latest commit information.



This library helps with a few things:

  • Querying
    • Provides helpers functions for retrieving entities when you don't need to specify their relationships to other entities and you only have one input, the db.
    • Includes a couple utility functions like ent-count and tempids
  • Tasks Makes it easier to create a database and add schema attributes. Assumes schemas are in resources/schemas
  • Data Loading Allows you to more succinctly express entities and relationships for loading into datomic. This is good for loading seed data, for example


In your project.clj, use

[com.flyingmachine/datomic-junk "0.2.1"]


Functions from the com.flyingmachine.datomic-junk namespace allow you to more succinctly express simple queries. Examples:

;; datomic - find the biebs
(q '[:find '?c :where [['?c :person/name "Justin Biebs"]]] db)

;; datomic-junk
(one db [:person/name "Justin Biebs"])

;; datomic - find all people
(q '[:find '?c :where [['?c :person/name]]] db)
;; datomic-junk
(all db :person/name)

;; datomic - find people with name X
(q '[:find '?c :where [['?c :person/name "X"]]] db)
;; datomic-junk
(all db [:person/name "X"])

;; Each datomic-junk function can take multiple conditions
(all db [:person/birth-year 1955] [:favorite/color "burgundy"])

Also check out the tests for more examples.


In the com.flyingmachine.datomic-junk.tasks namespace:

install-schemas takes a seq of schema names like

["topic" "post" "tag"]

and reads the corresponding .edn files under resources/schemas. It then runs the transactions in those files if they haven't been run already.

recreate recreates a database.

Data Loading

Functions in the com.flyingmachine.datomic-junk.load namespace allow you to more succinctly write entity and relationship data for loading into datomic. For example, say you have a simple forum schema with tags, topics, and posts, like this (truncated):

[;; tag
 {:db/ident :tag/name
  :db/valueType :db.type/string}
 ;; topic
 {:db/ident :topic/title
  :db/valueType :db.type/string}
 {:db/ident :topic/tags
  :db/valueType :db.type/ref
  :db/cardinality :db.cardinality/many}

 ;; post
 {:db/ident :post/content
  :db/valueType :db.type/string}
 {:db/ident :post/topic
  :db/valueType :db.type/ref}]

You could create entities using this data:

[{:topic/title "Were-Simmons?"
  :topic/tags [{:tag/name "health"}
               {:tag/name "emergency"}]}
 [{:post/topic {:topic/title "Were-Simmons?"}}
  {:post/content "I was bit by a were-Simmons. Is there any hope for me?"}
  {:post/content "No, sorry. You will turn into Richard Simmons on the full moon."}]]

This would create one topic with two tags and two posts. The first element of the vector is a map, which tells datomic-junk to create an entity. Its :topic/tags attribute is a vector, and datomic-junk creates those entities and associates them with the topic.

The second element is a vector. This tells datomic-junk to merge the first map of the vector ({:post/topic {:topic/title "Were-Simmons?"}}) with all of the maps that follow. It's as if you wrote this:

[{:topic/title "Were-Simmons?"
  :topic/tags [{:tag/name "health"}
               {:tag/name "emergency"}]}
 {:post/content "I was bit by a were-Simmons. Is there any hope for me?"
  :post/topic {:topic/title "Were-Simmons?"}}
 {:post/content "No, sorry. You will turn into Richard Simmons on the full moon."
  :post/topic {:topic/title "Were-Simmons?"}}]

When the value of a map key is itself a map, like {:post/topic {:topic/title "Were-Simmons?"}}, then datomic-junk uses the inner map as a query. In this way, each of the posts is associated with the topic that gets created.