Skip to content

Latest commit

 

History

History
157 lines (103 loc) · 4.49 KB

README.md

File metadata and controls

157 lines (103 loc) · 4.49 KB
Raw

porsas cljdoc badge

Nopea kuin sika pakkasella

Spike to see how fast we can go with both Clojure + JDBC & Async SQL. Highly Experimental.

Related dicsussion: https://clojureverse.org/t/next-jdbc-early-access/4091

Latest version

Clojars Project

Basics

porsas provides tools for precompiling the functions to convert database results into Clojure values. This enables basically Java-fast database queries using idiomatic Clojure.

SQL queries are executed against a Context, which caches compiled row transformation functions based on database metadata.

There are different Context implementations:

Currently, only eager query-one and query functions are supported.

Performance

At least an order of magnitude faster than clojure.java.jdbc, see the tests for more details.

Usage

JDBC

With defaults:

(require '[porsas.jdbc])

(def ctx (jdbc/context))

(jdbc/query-one ctx connection ["select * from fruit where appearance = ?" "red"])
; {:ID 1, :NAME "Apple", :APPEARANCE "red", :COST 59, :GRADE 87.0}

Returning maps with qualified keys:

(def ctx
  (jdbc/context
    {:key (jdbc/qualified-key)}))

(jdbc/query-one ctx connection ["select * from fruit where appearance = ?" "red"])
; #:FRUIT{:ID 1, :NAME "Apple", :APPEARANCE "red", :COST 59, :GRADE 87.0}

Generate Records for each unique Resultset, with lowercased keys. NOTE: this feature uses runtime code generation, so it doesn't work under GraalVM:

(def ctx
  (jdbc/context
    {:row (jdbc/rs->compiled-record)
     :key (jdbc/unqualified-key str/lower-case)}))

(jdbc/query-one ctx connection ["select * from fruit where appearance = ?" "red"])
; ; => #user.DBResult6208{:id 1, :name "Apple", :appearance "red", :cost 59, :grade 87.0}

Context can be omitted, bypassing all caching. Can be used when performance doesn't matter, e.g. when exploring in REPL:

(jdbc/query-one connection ["select * from fruit where appearance = ?" "red"])
; {:ID 1, :NAME "Apple", :APPEARANCE "red", :COST 59, :GRADE 87.0}

Async SQL

Uses non-blocking vertx-sql-client and can be used with libraries like Promesa and Manifold.

(require '[porsas.async :as async])

(def ctx (async/context))

;; define a pool
(def pool
  (async/pool
    {:uri "postgresql://localhost:5432/hello_world"
     :user "benchmarkdbuser"
     :password "benchmarkdbpass"
     :size 16}))

(-> (async/query-one ctx pool ["SELECT randomnumber from WORLD where id=$1" 1])
    (async/then :randomnumber)
    (async/then println))
; prints 504

A blocking call:

(-> (async/query-one ctx pool ["SELECT randomnumber from WORLD where id=$1" 1])
    (async/then :randomnumber)
    (deref))

With Promesa

(require '[promesa.core :as p])

(-> (pa/query-one ctx pool ["SELECT randomnumber from WORLD where id=$1" 1])
    (p/chain :randomnumber println))
; #<Promise[~]>
; printls 504

With Manifold

(require '[manifold.deferred :as d])

(-> (pa/query-one ctx pool ["SELECT randomnumber from WORLD where id=$1" 1])
    (d/chain :randomnumber println))
; << … >>
; printls 504

next.jdbc

Using porsas with :builder-fn option of next.jdbc:

(require '[porsas.next])
(require '[next.jdbc])

(def builder-fn (porsas.next/caching-row-builder))

(next.jdbc/execute-one! connection ["select * from fruit where appearance = ?" "red"] {:builder-fn builder-fn})
; #:FRUIT{:ID 1, :NAME "Apple", :APPEARANCE "red", :COST 59, :GRADE 87.0}

More info

There is #sql in Clojurians Slack for discussion & help.

Roadmap as issues.

License

Copyright © 2019 Metosin Oy

Distributed under the Eclipse Public License, the same as Clojure.