From Relational To Document

#tech4africa presentation

Transcript

1. From Relational to Document Wingman, MongoDB Norberto Leite #T4A
2. Some introductions … •  Norberto Leite •  Solutions Architect, MongoDB Inc •  Nice Guy •  @nleite
3. Did I mentioned that love being in South Africa? SUPER LOVING IT
7. No worries, not going to mention Pistorus !
8. Data Structures
9. Key -> Value •  One-dimensional storage •  Single value is a blob •  Query on key only •  No schema •  No updates just replaces Key Blob
10. Relational •  2-dimensional storage (tuples) •  Each field contains one value •  Query on any field •  Very structured schema (tables) •  In-place updates •  Normalization requires many tables, joins, indexes •  Poor data locality Primary Key
11. Document •  N-dimensional storage •  Each field can contain 0,1, many or embedded values •  Query on any field & level •  Flexible schema •  Inline updates •  Embedding related data has optimal data locality, requires fewer indexes, has better performance _id
12. Relational
13. Relational Track Record How many of you learn SQL at school?
15. http://www.orafaq.com/wiki/images/thumb/f/f2/Sqlplus.jpg/400px-Sqlplus.jpg
16. http://xkcd.com/327/
19. (3months later) Great! I know what I want Let’s Get The Show On The Road
20. Entity Mapping Normalization 6th Normal Form Path to Normalization
21. Great we have a correct schema! Now Let’s Make It Work!!!
22. Break Joins Removing Transactions Remove Foreign Keys Multi Field Tables BLOBS of XML What the hell is this data doing in RDBMS? Path to DeNormalization (Sanity)
24. http://geekandpoke.typepad.com/.a/6a00d8341d3df553ef0162fd7fe2e6970d-pi
25. My Database Professor at University “You normalize it till is correct, then you denormalize till it works”
26. Documents
27. So we are storing pdf’s right? …
28. { "_id" : 451152, "amenity" : "pub", "name" : "The Dignity", "addr:housenumber" : "363", "addr:street" : "Regents Park Road", "addr:city" : "London", "addr:postcode" : "N3 1DH", "toilets" : "yes", "toilets:access" : "customers", "location" : { "type" : "Point", "coordinates" : [-0.1945732, 51.6008172] } } Example of Document
29. Relational MongoDB { ! first_name: ‘Paul’,! surname: ‘Miller’,! city: ‘London’,! location: [45.123,47.232],! cars: [ ! { model: ‘Bentley’,! year: 1973,! value: 100000, … },! { model: ‘Rolls Royce’,! year: 1965,! value: 330000, … }! }! }! Document Data Model
30. Document Model Benefits •  Agility and Flexibility –  Data models can evolve easily –  Projects can adapt to changes quickly •  Intuitive, natural data representation –  Developers are more productive –  Many types of applications are a good fit •  Reduces the need for joins, disks seeks –  Programing is more simple –  Performance can be delivered at scale
31. class Pearson{ String name; String passport; int age; String []countries_visited; Nationality nationality; } class Nationality{ String country; String name; } Seamless Integration
32. { _id: ObjectId(), name: “Norberto”, passport: “S3CR3T”, age: 30, countries_visited: [‘Brasil’, ‘Portugal’, ‘Spain’, ‘Italy’, ‘Israel’ …], nationality: { country: ‘Portugal’, name: ‘Portuguese’ } } To Documents
33. Developers More Productive
34. Much More Productive
35. Way More Productive
36. From Relational to Document
37. Step 1 Start defining your App
39. Step 2 Start building your code
41. Step 3 Right, you want to add this feature!
42. Most logical comment I’ve ever heard “Requirements are like ice skating, only works when they are frozen”
43. This field should be VARCHAR2 Need FK on this other Don’t need this relation after all
44. Step 4 Migrations and Stuff
48. Step 5 DeNormalize
50. Denormalize •  Break Joins •  Remove Transactions •  Merge Tables •  Merge Fields •  Temporary Tables •  Transient Data Structures
51. Your database morphs into Excell Spreadsheets
53. Step 6 You know what, that MongoDB thing sounds pretty cool!
55. Demo
56. Obrigado norberto@mongodb.com @nleite #T4A