Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Postgresql + Ruby = :heart:
Search
Yannick Schutz
April 08, 2014
Programming
4
320
Postgresql + Ruby = :heart:
How you could use postgresql in ways you didn't imagine when using it as a simple datastore.
Yannick Schutz
April 08, 2014
Tweet
Share
More Decks by Yannick Schutz
See All by Yannick Schutz
All those bots are gonna steal your job
ys
1
1.2k
Ruby loves Postgres
ys
1
160
RubyMotion - Apéro Ruby Paris 05-10-2012
ys
1
3k
Other Decks in Programming
See All in Programming
AIの弱点、やっぱりプログラミングは人間が(も)勉強しよう / YAPC AI and Programming
kishida
10
5.2k
無秩序からの脱却 / Emergence from chaos
nrslib
1
6.4k
Promise.tryで実現する新しいエラーハンドリング New error handling with Promise try
bicstone
3
750
Duke on CRaC with Jakarta EE
ivargrimstad
0
150
Eloquentを使ってどこまでコードの治安を保てるのか?を新人が考察してみた
itokoh0405
0
3.2k
OSS開発者の憂鬱
yusukebe
12
5.5k
PHPライセンス変更の議論を通じて学ぶOSSライセンスの基礎
matsuo_atsushi
0
170
物流DXを支える“意味”の設計:セマンティックレイヤーとAIで挑むデータ基盤/登壇資料(飯塚 大地)
hacobu
PRO
0
110
Bakuraku E2E Scenario Test System Architecture #bakuraku_qa_study
teyamagu
PRO
0
780
TypeScriptで設計する 堅牢さとUXを両立した非同期ワークフローの実現
moeka__c
0
380
チーム開発の “地ならし"
konifar
8
5.7k
AI時代もSEOを頑張っている話
shirahama_x
0
130
Featured
See All Featured
Faster Mobile Websites
deanohume
310
31k
Exploring the Power of Turbo Streams & Action Cable | RailsConf2023
kevinliebholz
36
6.1k
Raft: Consensus for Rubyists
vanstee
140
7.2k
Side Projects
sachag
455
43k
Git: the NoSQL Database
bkeepers
PRO
432
66k
What's in a price? How to price your products and services
michaelherold
246
12k
Art, The Web, and Tiny UX
lynnandtonic
303
21k
The Pragmatic Product Professional
lauravandoore
36
7k
Automating Front-end Workflow
addyosmani
1371
200k
Optimising Largest Contentful Paint
csswizardry
37
3.5k
Building Flexible Design Systems
yeseniaperezcruz
329
39k
Mobile First: as difficult as doing things right
swwweet
225
10k
Transcript
Ruby+ Postgresql=♥♥♥♥
Hello, I’m Yannick
None
None
PostgreSQL is more than a datastore
PostgreSQL is a relational database
PostgreSQL is webscale
PostgreSQL is webscale
PostgreSQL is open-source
PostgreSQL is maintained
PostgreSQL has a great community
PostgreSQL has new awesome features coming!
PostgreSQL will do your coffee
Let me show you it’s POWER
PROBLEM
You fetch something from an API call. Like that Oauth
call to twitter.
a) You add a new cache layer like Redis b)
You add a table with all the fields c) You add fields to an existing model d) Something else?
a) You add a new cache layer like Redis b)
You add a table with all the fields c) You add fields to an existing model d) Something else?
a) You add a new cache layer like Redis b)
You add a table with all the fields c) You add fields to an existing model d) Something else?
a) You add a new cache layer like Redis b)
You add a table with all the fields c) You add fields to an existing model d) Something else?
a) You add a new cache layer like Redis b)
You add a table with all the fields c) You add fields to an existing model d) Something else?
JSON
class AddJsonPayloadToTwitterAuth < ActiveRecord::Migration def change add_column(:twitter_auths, :json_payload, :json) end
end
PostgreSQL learn you japanese
PROBLEM
You add new user settings every two weeks and needs
new avatar sizes every three.
a) You store that in Redis b) You add a
table with all the fields c) You add fields to the existing model d) Something else?
a) You store that in Redis b) You add a
table with all the fields c) You add fields to the existing model d) Something else?
a) You store that in Redis b) You add a
table with all the fields c) You add fields to the existing model d) Something else?
a) You store that in Redis b) You add a
table with all the fields c) You add fields to the existing model d) Something else?
a) You store that in Redis b) You add a
table with all the fields c) You add fields to the existing model d) Something else?
HSTORE
add_column :users, :settings, :hstore ! class User < ActiveRecord::Base #
This exposes accessors. user.wants_push = false store_accessor :settings, :wants_push, :wants_mails, :auto_save store_accessor :avatars, :large, :medium, :small end
PostgreSQL is not mysql
PROBLEM
You have a really complex query that is blazing fast
but when you come back a week after you’re like ‘WTF is that!’
a) You keep it unreadable b) You change the behavior
c) Something else?
a) You keep it unreadable b) You change the behavior
c) Something else?
a) You keep it unreadable b) You change the behavior
c) Something else?
a) You keep it unreadable b) You change the behavior
c) Something else?
CTE AKA WITH clause
def bad_users_with_karma User.find_by_sql(query) end ! def query <<-SQL WITH bad_users
AS (SELECT * FROM users WHERE bad = true) SELECT * FROM bad_users WHERE karma > #{karma_limit} SQL end
VIEWS THANKS PSQL
PostgreSQL believes you can fly
PROBLEM
You really need to add all this complexity that makes
that query run in seconds…
a) You use a complex query b) You denormalize in
MongoDB c) Something else?
a) You use a complex query b) You denormalize in
MongoDB c) Something else?
a) You use a complex query b) You denormalize in
MongoDB c) Something else?
a) You use a complex query b) You denormalize in
MongoDB c) Something else?
VIEWS
CREATE MATERIALIZED VIEW bad_users AS SELECT * FROM users WHERE
bad = false; ! class BadUser < User def readonly? true end end
PostgreSQL might contains kittens and unicorns
PROBLEM
You need to search your users by username, name and
email.
a) You add that elastic search b) You select in
ruby c) Something else?
a) You add that elastic search b) You select in
ruby c) Something else?
a) You add that elastic search b) You select in
ruby c) Something else?
a) You add that elastic search b) You select in
ruby c) Something else?
SEARCH
# Using textacular ! User.basic_search(“yannick”) User.basic_search(name: “yan:*”) User.fuzzy_search(“yannick”) # uses
trigrams ! create index on users using gin(to_tsvector('english', name));
PostgreSQL comes in multiple flavors and colors
Now, you know that the elephant database is more than
a datastore!
Thanks!
Questions?
Questions?