begriffs

import Control.Monad
import Data.Int
import Data.Monoid
import Data.Functor.Contravariant

import Hasql.Connection
import Hasql.Session
import Hasql.Query
import qualified Hasql.Decoders as HD
import qualified Hasql.Encoders as HE

begriffs

createReadStatement
  asCsvF
  asJsonF
  asJsonSingleF
  limitF
  locationF

createWriteStatement
  asCsvF
  asJsonF

begriffs

-- JSON FESTIVAL

WITH pg_source AS
  (SELECT "public"."festival".* FROM "public"."festival")
SELECT
  (SELECT pg_catalog.Count(1) FROM "public"."festival")       AS total_result_set,
  pg_catalog.Count(t)                                         AS page_total,
  NULL                                                        AS header,
  Array_to_json(Array_agg(Row_to_json(t)))::character VARYING AS body
FROM

begriffs

#lang racket
(require quickcheck)

(define my_gt
  (lambda (n m)
    (cond
      ((zero? n) #f)
      ((zero? m) #t)
      (else (my_gt (sub1 n) (sub1 m))))))

begriffs

curl -s https://app.workshopcafe.com/seats/taken_and_unreservable_seats | json | grep \\b7[4-6]

begriffs

Answers to objections about putting logic in the db

• The DB lacks version control, and a team cannot collaborate. There is no consistent way to evolve the schema.
  • Keep all db code in version control as test-driven migrations
  • Don't change the db except through migrations
  • Use a tool like Sqitch to apply the migrations and keep track of migration history using an auxiliary set of tables in the db
  • Write tests for views and functions with plpgunit
  • Deployment is easier because changing a backend means changing the database. Just point sqitch at the staging/production db and apply the migration(s) same as you would locally.
• Central computation in a database can't scale
  • Start with RDS which allows you to scale pretty far

begriffs

2015-12-28T03:32:53.153150+00:00 heroku[router]: at=info method=OPTIONS path="/speakers?id=eq.140" host=postgrest.herokuapp.com request_id=dcdd0fb5-98f3-45ae-a769-4d6be6a4b6c8 fwd="73.233.161.119" dyno=web.1 connect=1ms service=1ms status=200 bytes=427
2015-12-28T03:32:53.144669+00:00 heroku[router]: at=info method=OPTIONS path="/speakers?id=eq.140" host=postgrest.herokuapp.com request_id=bfd29fc2-942e-456a-aba1-74346dbb1a23 fwd="73.233.161.119" dyno=web.1 connect=1ms service=16ms status=200 bytes=427
2015-12-28T03:32:53.132313+00:00 heroku[router]: at=info method=OPTIONS path="/speakers?id=eq.140" host=postgrest.herokuapp.com request_id=2121268b-397c-4fe9-90e9-b842c1e2c668 fwd="73.233.161.119" dyno=web.1 connect=1ms service=1ms status=200 bytes=427
2015-12-28T03:32:53.133157+00:00 app[web.1]: 10.180.57.199 - - [28/Dec/2015:03:32:53 +0000] "OPTIONS /speakers HTTP/1.1" 200 - "http://timwis.com/handsontable-postgrest/" "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_11_2) AppleWebKit/537.36 (KHTML, like Gecko) Chrom

begriffs

primesTo :: Int -> [Int]
primesTo n = removeHeadMultiples [2..n]
 where
  removeHeadMultiples same@(p:xs)
    | p*p > n   = same -- short circuit factors greater than sqrt(n)
    | otherwise = p : removeHeadMultiples [x | x <- xs, x `rem` p > 0]

begriffs

#%RAML 1.0
title: PostgREST Demo
version: v1
baseUri: https://postgrest.herokuapp.com/
types:
  Sessions:
    type: object
    properties:
      id: integer
      speaker_id: integer

begriffs

	Sun Nov 22 15:25 2015 Time and Allocation Profiling Report  (Final)

	   postgrest +RTS -p -L500 -RTS -- postgresql://postgrest_test:@localhost:5432/postgrest_test -a postgrest_test -s test

	total time  =        8.31 secs   (8311 ticks @ 1000 us, 1 processor)
	total alloc = 40,138,127,800 bytes  (excludes profiling overheads)

COST CENTRE MODULE                     %time %alloc

MAIN        MAIN                        95.9   98.4