Postgres SQL is a plain old SQL DB. Very powerful and very consistent, in some case, project needs an ACID database, but schemaless… With JSON support, postgres is a very interesting tool to provide ACID and some very interesting function (time management, localisation function and data types...) and the schemaless noSQL point of view with json and indexed json. This talk show some great usage and some insigth to build some great application with postgres.

1. WHY POSTGRES
SQL DESERVE
NOSQL FAN
RESPECT
QUENTIN ADAM
@WAXZCE

2. Quentin ADAM from the Clever Cloud
@waxzce on twitter – github- soundcloud – instagram ….
WHO AM I ?

3. MY DAY TO DAY WORK :
CLEVER CLOUD, THE IT
AUTOMATION COMPANY

4. And learn a lot of things about your code, apps, and good/bad design…
KEEP YOUR APPS ONLINE. MADE WITH
NODE.JS, SCALA, JAVA, RUBY, PHP,
PYTHON, GO…

5. AND LEARN A LOT OF THINGS ABOUT
YOUR CODE, APPS, AND GOOD/BAD
DESIGN…

6. NOSQL WORLD

8. SOME OF THE COOL THINGS
• ACID or not
• Schemaless or not
• Concurrent writing or not
• Consistent or eventually consistent
• Manage mass of data…
• HA or not
• …

9. THINK ABOUT THE DATA
STORAGE

10. LET’S EXAMPLE

11. PLAYLIST
MANAGEMENT APP,
MULTIPLE
BACKEND,
PLAYLIST HISTORY
TO DO

12. ENTITIES
• user management
• Playlists
• Tracks history (deezer, spotify, soundcloud)

13. TRACK DATA IS
HETEROGENEOUS

14. Use nosql
WE NEED SCHEMALESS

15. ARBITRARY DATA DEPTH

16. EXTERNAL DATA IS ALREADY JSON

17. “I NEED
UNIQUE
USER ON
EMAIL”

18. CONSTRAINTS

19. “WHICH
TRACKS ARE
PLAYED AT
8PM TO 9PM
WEEK DAYS?”

20. COMPLEX DATA
EXPLORATORY IS HARD

21. “… IN PARIS
AREA”

22. COMPLEX DATA
EXPLORATORY IS HARD

23. “… BY MALE
USERS”

24. NEED SOME RELATIONSHIPS IN YOUR
DATA STORE?

25. “…ORDER
BY..”
“…UI NEED
PAGINATION
…”

26. WE
ALREADY
KNOW ALL
OF THESE

27. POSTGRESQL

28. LET’S CREATE OUR TRACKS
TABLE

29. UUID AS
PRIMARY
KEY
& NO
SERIAL

30. “WE HAVE 2000000 USERS
ON OUR PRODUCT”
http://www.my-platform.com/users-creation-confirm/42
?valid_key=jgqsdkfhqksdjfhkjsdhf2342345

31. USING IT TO COUNT THINGS
>_<

32. SAME ID FOR VARIOUS ENTITIES
ON YOUR DATABASE <_>

33. Generate it on the code level
WHAT IS MY ID?

34. PUT A
NAME ON
THINGS
COMPUTER SCIENCE PROBLEM

35. https://www.clever-cloud.com/blog/engineering/2015/05/20/why-auto-increment-is-a-
terrible-idea/
BLOG ARTICLE BY THE AWESOME
@CLEMENTD (CLEVER CLOUD CTO)

36. LET’S CREATE OUR TRACKS
TABLE

37. USE ENUM

38. USE TIMESTAMPTZ

39. LET’S CREATE OUR TRACKS
TABLE

40. JSON DATA ON YOUR DB

41. NATIVE JSON & JSONB TYPE

42. LET’S CREATE OUR TRACKS
TABLE

43. THEN EMBRACE SQL POWER

44. *WITH REAL DATABASE LIKE
POSTGRE SQL

45. REQUEST JSON

46. OR TEST EXISTENCE

47. USE SQL FUNCTION ON IT

48. Array length, data manipulation, filters, jsonpath, prettyfier
PLENTY OF CAPACITY FOR JSON

49. TEMPORAL FUNCTION WITH
PGSQL

50. COUNT(*) ON NOSQL?

51. GEO BASED REQUEST?

52. INDEX

53. INDEX

54. TRANSACTIONS

55. ACID

56. BOOTSTRAPPING?

57. “MY PLAIN
OLD SQL IS
SO
POWERFUL?
”

58. CHOOSE DB (AND VERSION)

59. KILL THE ORM

60. JDBC DRIVERS

61. WORKAROUND

62. WORKAROUND

63. MANAGE IT LIKE REGULAR
JSON IN YOUR CODE

64. FREE YOUR MIND ABOUT
DATASTORE + DRIVER ACCESS

65. “SO PGSQL
IS BETTER
THAN
MONGO?”

66. ACID
PRO/CON

67. INSTANCE FOR ONE
ORGANIZATION
ACID
• Atomicity
• Consistency
• Isolation
• Durability
Powerful
data
management
• Transaction
• User
management
• One above one
Take advantage of ACID
database

68. BUT IT DESIGNS A
BOTTLENECK

69. EXAMPLE : E-SHOP ON
CLASSIC MODE

70. EXAMPLE : E-SHOP ON
CLASSIC MODE
User A buy a
hdd
Database Transaction :
• Stock management
• Order management
• Invoice generation
• Customer Account reward
• …
Transaction
user A is
processed
Stock &
Order are just
perfectly
synchronize

71. EXAMPLE : E-SHOP ON
CLASSIC MODE
User A buy a
hdd
Database Transaction :
• Stock management
• Order management
• Invoice generation
• Customer Account reward
• …
Transaction
user A is
processed
Stock &
Order are just
perfectly
synchronize
User B buy a
hdd Transaction
user B is
processed
then

72. EXAMPLE : E-SHOP ON
MULTI-TENANT MODE
i.e. : Multiple shop of various sellers on the same instance

73. EXAMPLE : E-SHOP ON
MULTI-TENANT MODE
User A buy a
hdd on seller A
Database Transaction :
• Stock management
• Order management
• Invoice generation
• Customer Account reward
• …
Transaction
user A is
processed
Stock &
Order are just
perfectly
synchronize
User B buy a
book on seller B Transaction
user B is
processed
then

74. IS YOUR
DATASTORE
WORK FOR
MULTI
TENANCY ?

75. CONCURRENT INSERT

76. SHARDING & REPLICA

77. HIGH AVAILABILITY

78. CHOOSE A
DATASTORE IS
REAL
ARCHITECTURE
JOB ;-)

79. PGSQL OPS
TRICK FOR
GROWING
DB

80. AM I LOCK ON ONLY ONE NODE?

81. Using live replication
MASTER/SLAVE

82. SPLIT REQUEST ON YOUR CODE

83. Able to make requests navigation
OR USE PGPOOL

84. BI for example, keep production DB stable
USE REPLICA FOR SPECIFIC PURPOSE

85. Use proxy for your DB, as PG bouncer
DB STABILITY IS DRIVEN BY
PREDICTABILITY

86. http://dalibo.github.io/powa/
GET METRICS, LEARN, IMPROVE

87. TRY IT
BEFORE :-)

88. FREE DEV
PLAN ON
CLEVER
CLOUD

89. ONE CLICK LAUNCH

90. NOSQL = NOT ONLY SQL

91. I’m @waxzce on twitter
Take some time to try
THX FOR LISTENING
& QUESTIONS TIME
Clever Cloud coupon : rigadevday16