CouchDB Talk, Codestock 2012

Transcript

1. Getting Started with CouchDB and .NET

2. Who is this Guy? Josh Bush digitalbush.com freshbrewedcode.com fireflylogic.com @digitalBush

3. CouchDB • Document Database • Open Source • Written in

   Erlang • RESTful HTTP API • JSON • Query via Map/Reduce • Bi-directional replication

4. Right Tool, Right Job. • This is not a SQL

   hate fest. • CouchDB is not a silver bullet. • Use the tool that works best for the problem you are trying to solve.

5. NoSQL Describes a broad range of database technologies which are

   non-relational. Focused towards solving specific problems which are not suited for relational storage.

6. Documents • Just a collection of key value pairs. •

   Represented as JSON. • No need to store things separately which naturally go together. • Can also have attachments. • Retrieved via key.

7. JSON Document { “Name”:”Document”, “Description”:”Collection of key/value pairs”, “IsAwesome”:true, “Value-Types”:[“Number”,”String”,”Boolean”,”Array”,”Object”],

   “Volume”: 11, “Has-The-Following-Default-Properties”:{ “_id”: “uniquely identifies the document”, “_rev”: “version of the document” } }

8. Example Order #7 Burger French Fries Coke Bun Type Meat

   Type Toppings Lettuce Tomato Mustard Onion Pickles

9. Relational Database Order #7 Burger French Fries Coke Bun Type

   Meat Type Toppings Lettuce Tomato Mustard Onion Pickles Order Table

10. Relational Database Order #7 Burger French Fries Coke Bun Type

    Meat Type Toppings Lettuce Tomato Mustard Onion Pickles Order Table OrderItems Table

11. Relational Database Order #7 Burger French Fries Coke Bun Type

    Meat Type Toppings Letuce Tomato Mustard Onion Pickle Order Table Don’t forget the foreign key! OrderItems Table

12. Relational Database Order #7 Burger French Fries Coke Bun Type

    Meat Type Toppings Lettuce Tomato Mustard Onion Pickles Order Table Don’t forget the foreign key! OrderItems Table BurgerItem Table

13. Relational Database Order #7 Burger French Fries Coke Bun Type

    Meat Type Toppings Lettuce Tomato Mustard Onion Pickles Order Table Don’t forget the foreign key! OrderItems Table BurgerItem Table Another Foreign Key (One to One)

14. Relational Database Order #7 Burger French Fries Coke Bun Type

    Meat Type Toppings Lettuce Tomato Mustard Onion Pickles Order Table Don’t forget the foreign key! OrderItems Table BurgerItem Table Another Foreign Key (One to One) Toppings Table

15. Relational Database Order #7 Burger French Fries Coke Bun Type

    Meat Type Toppings Lettuce Tomato Mustard Onion Pickles Order Table Don’t forget the foreign key! OrderItems Table BurgerItem Table Another Foreign Key (One to One) Toppings Table MOAR Foreign Key!

16. Document Database Order #7 Burger French Fries Coke Bun Type

    Meat Type Toppings Lettuce Tomato Mustard Onion Pickles Store in one document

17. Document Database { “_id”: 7, “Items”: [ { “Type”: “Burger”,

    “BunType”: “Sesame Seed”, “MeatType”: “1/4 lb”, “Toppings”: [“Lettuce”,”Tomato”,”Mustard”,”Onion”,”Pickles”] }, {“Type”: “French Fries”}, {“Type”: “Coke”} ] } Represented as JSON

18. Time to Make a Mess

19. Document API Don’t be scared, it’s just HTTP & JSON

20. API • Create • POST http://{host}/{database} • PUT http://{host}/{database}/{id} •

    Read • GET http://{host}/{database}/{id} • Update • PUT http://{host}/{database}/{id} • Delete • DELETE http://{host}/{database}/{id}?rev={rev}

21. (Finally, time to see some code)

22. That Was Fun? My demos show it the “hard” way.

    You have other options: • RestSharp (http://restsharp.org/) • New HttpClient in .NET 4.5 (or NuGet) • Relax (https://github.com/arobson/Relax) • LoveSeat (https://github.com/soitgoes/LoveSeat)

23. View API Don’t be scared, it’s just HTTP & JSON

24. Map/Reduce • Used to create views on the data. •

    Roughly equivalent to a materialized view(snapshot) with indexing (sort of). • Useful for querying/some reporting.

25. Map • Takes data in and transforms it into something

    else. • Output is key/value pairs. • Keys can be complex types. • Like a select in LINQ.

26. Reduce • Takes in a set of intermediate values and

    combines them into a single value. • Reduce needs to be able to accept results from the map function AND the reduce function itself. • Like an .Aggregate() in LINQ

27. Map/Reduce in LINQ var test=new [] {"1","2","3”}; var mapped =

    test.Select(Int32.Parse); var reduced = mapped.Aggregate((sum,i)=>sum+=i);

28. Map/Reduce in CouchDB "map": "function(doc){ emit(doc._id, parseInt(doc.value)); }", "reduce": "function(keys,values)

    { return sum(values); }"

29. View API • Read • GET http://{host}/{database}/_design/{doc}/_view/{name} • Map Only

    • ?reduce=false • Key • Specific - ?key={key} • Range - ?startkey={lower bound}&endkey={upper bound} • Grouping • ?group_level={n} • Include Documents • ?include_docs=true

30. Powershell too? Why not.

31. More Demo? Wikipedia/CouchDB demo if we have time.

32. Demos From Talk https://github.com/digitalBush/Couchipedia https://github.com/digitalBush/couch-samples