CouchDB Vs MongoDB

Document
{
"day": [ 2010, 01, 23 ],
"products": {
"apple": {
"price": 10

Key ->
"quantity": 6
},
"kiwi": {
"price": 20
"quantity": 2
}
},
"checkout": 100
}

Couchdb Mongodb

Data Model Document-Oriented (JSON) Document-Oriented (BSON)

Interface HTTP/REST Custom protocol over TCP/IP

Object Database contains Collections
Database contains Documents
Storage Collections contains Documents

Map/Reduce (javascript) creating
Query Map/Reduce (javascript + others)
Collections + Object-Based query
Method creating Views + Range queries
language
Master-Master with custom
Replication Master-Slave
conﬂict resolution functions

MVCC (Multi Version Concurrency
Concurrency Update in-place
Control)

Written In Erlang C++

Example: Tickets

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100123, "day": 20100123, "day": 20100123,
"checkout": 100 "checkout": 42 "checkout": 215 "checkout": 73
} } } }

Sum(checkout)?

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100123, "day": 20100123, "day": 20100123,
} } } }

Map: emit(checkout)

100 42 215 73

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100123, "day": 20100123, "day": 20100123,
} } } }

Reduce: sum(checkouts)

142 288

100 42 215 73

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100123, "day": 20100123, "day": 20100123,
} } } }


430

142 288

100 42 215 73

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100123, "day": 20100123, "day": 20100123,
} } } }

Reduce must be associative

reduce( 100 42 215 73 ) == 430

Must be equal to
reduce(

reduce( 100 42 ) == 142

reduce( 215 73 ) == 288

) == 430

SELECT
SUM(checkout)
FROM ticket
?!?!?!?

Inherently distributed

430

142 288

100 42 215 73

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100123, "day": 20100123, "day": 20100123,
} } } }

Logaritmic Update

430

142 288

100 42 215 73

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100123, "day": 20100123, "day": 20100123,
} } } }

Logaritmic Update

430

142 288

100 42 210 73

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100123, "day": 20100123, "day": 20100123,
} } } }

Logaritmic Update

430

142 283

100 42 210 73

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100123, "day": 20100123, "day": 20100123,
} } } }

Logaritmic Update

425

142 283

100 42 210 73

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100123, "day": 20100123, "day": 20100123,
} } } }

Sum(checkout)
# START SERVER
$ ~/opt/mongodb-1.3.0/bin/mongod
--dbpath=./db/mongodb.01/
--logpath=./log/mongodb.01
--port 30001

# START SHELL
$ ~/opt/mongodb-1.3.0/bin/mongo localhost:30001
connecting to: localhost:30001/test
type "help" for help
> show dbs
admin
local

Sum(checkout)
> use checkout
switched to db checkout

> db.tickets.save({ "_id": 1, "day": 20100123, "checkout": 100 })

> db.tickets.count()
4

> db.tickets.find()
{ "_id" : 1, "day" : 20100123, "checkout" : 100 }
...

> db.tickets.find({ "_id": 1 })
{ "_id" : 1, "day" : 20100123, "checkout" : 100 }

Sum(checkout)
> var map = function() {
... emit(null, this.checkout)
... }

> var reduce = function(key, values) {
... var sum = 0
... for (var index in values) sum += values[index]
... return sum
... }

Sum(checkout)
Temporary Collection
> sumOfCheckouts = db.tickets.mapReduce(map, reduce)
{
"result" : "tmp.mr.mapreduce_1263717818_4",
"timeMillis" : 8,
"counts" : { "input" : 4, "emit" : 4, "output" : 1 },
"ok" : 1
}

> db.getCollectionNames()
[
"tickets",
"tmp.mr.mapreduce_1263717818_4",
]

> db[sumOfCheckouts.result].find()
{ "_id" : null, "value" : 430 }

Sum(checkout)
Persistent Collection
> db.tickets.mapReduce(map, reduce, { “out”: “sumOfCheckouts” })

> db.getCollectionNames()
[
“sumOfCheckouts”,
"tickets",
"tmp.mr.mapreduce_1263717818_4"
]

> db.sumOfCheckouts.find()
{ "_id" : null, "value" : 430 }

> db.sumOfCheckouts.findOne().value
430

Sum(checkout)
Reduce by Group
# GROUP AS MAP/REDUCE ALTERNATIVE

> db.tickets.group({
... "initial": { "sum": 0 },
... "reduce": function(ticket, checkouts) {
...... checkouts.sum += ticket.checkout
...... }
... })
[ { "sum" : 430 } ]

Sum(checkout) Group By day

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100124, "day": 20100123, "day": 20100124,
} } } }

Map: emit(day,checkout)

“20100123”:100 “20100124”:42 “20100123”:215 “20100124”:73

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100124, "day": 20100123, "day": 20100124,
} } } }


“20100123”:315

“20100123”:100 “20100124”:42 “20100123”:215 “20100124”:73

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100124, "day": 20100123, "day": 20100124,
} } } }


“20100123”:315 “20100124”:115

“20100123”:100 “20100124”:42 “20100123”:215 “20100124”:73

{ { { {
"id": 1, "id": 2, "id": 3, "id": 4,
"day": 20100123, "day": 20100124, "day": 20100123, "day": 20100124,
} } } }

Design Documents are
Documents

Non trivial Map:
Calculate Checkout

Structured Keys and
Group Levels

Sum(Checkout) by day
Update In-Place
> db.tickets.update({ "_id": 1 }, {
... $set: { "products": {
...... "apple": { "quantity": 5, "price": 10 },
...... "kiwi": { "quantity": 2, "price": 25 }
...... }
... },
... $unset: { "checkout": 1 }
... })

> db.tickets.find()
{ "_id" : 1, "day" : 20100123, "products" : {
"apple" : { "quantity" : 5, "price" : 10 },
"kiwi" : { "quantity" : 2, "price" : 25 }
}}
{ "_id" : 2, "day" : 20100123, "checkout" : 42 }
{ "_id" : 3, "day" : 20100123, "checkout" : 215 }
{ "_id" : 4, "day" : 20100123, "checkout" : 73 }

Calculate Checkout
> db.tickets.find()
{ "_id" : 1, "day" : 20100123, "products" : {
"apple" : { "quantity" : 5, "price" : 10 },
"kiwi" : { "quantity" : 2, "price" : 25 } } }

{ "_id" : 2, "day" : 20100124, "products" : {
"banana" : { "quantity" : 2, "price" : 20 } } }

{ "_id" : 3, "day" : 20100123, "products" : {
"kiwi" : { "quantity" : 4, "price" : 25 },
"babana" : { "quantity" : 5, "price" : 20 },
"lemon" : { "quantity" : 3, "price" : 5 } } }

{ "_id" : 4, "day" : 20100124, "products" : {
"kiwi" : { "quantity" : 2, "price" : 25 },
"babana" : { "quantity" : 1, "price" : 20 } } }

Calculate Checkout
... var checkout = 0
... for (var name in this.products) {
...... var product = this.products[name]
...... checkout += product.quantity * product.price
...... }
... emit(this.day, checkout)
}

... var sum = 0
... return sum
}

Calculate Checkout
> db.tickets.mapReduce(map, reduce, { "out": "sumOfCheckouts" })

{ "_id" : 20100123, "value" : 315 }
{ "_id" : 20100124, "value" : 110 }

Data Normalization
> db.tickets.find()
{ "_id" : 1, "day" : 20100123, "products" : {
"apple" : 5, "kiwi" : 2 } }
{ "_id" : 2, "day" : 20100124, "products" : {
"banana" : 2 } }
{ "_id" : 3, "day" : 20100123, "products" : {
"kiwi" : 4, "banana" : 5, "lemon" : 3 } }
{ "_id" : 4, "day" : 20100124, "products" : {
"kiwi" : 2, "banana" : 1 } }

> db.product.find()
{ "_id" : "apple", "price" : 10 }
{ "_id" : "kiwi", "price" : 25 }
{ "_id" : "banana", "price" : 20 }
{ "_id" : "lemon", "price" : 5 }

Data Normalization
... var checkout = 0
... for (var name in this.products) {
...... var quantity = this.products[name]
...... var price = db.product.findOne({ "_id": name }).price
...... checkout += quantity * price
...... }
... emit(this.day, checkout)
}

... var sum = 0
... return sum
}

Data Normalization
> db.tickets.mapReduce(map, reduce, { "out": "sumOfCheckouts" })

{ "_id" : 20100123, "value" : 315 }
{ "_id" : 20100124, "value" : 110 }

Count of unique
elements?
> db.view.find();
{ "user" : "001", "page" : "example.com/001", "time" : 2 }

# USER NAVIGATION SURVEY = FOR EACH USER
# NUMBER OF UNIQUE PAGES
# AVERAGE TIME ON A PAGE

Count of unique
elements?
... var accumulator = {
...... "numberOfViews": 1,
...... "visitedPages": {},
...... "totalTime": 0
...... };

... accumulator["visitedPages"][this.page] = 1
... accumulator["totalTime"] += this.time
... emit(this.user, accumulator)
}

Count of unique
elements?
# EASY TO DEBUG

> var aUser = db.view.findOne({ "user": "001" })

> var emit = function(id, value) { print(tojson(value)) }

> map.call(aUser)
{
"numberOfViews" : 1,
"visitedPages" : {
"example.com/001" : 1
},
"totalTime" : 2
}

Count of unique
elements?
...... "visitedPages": {},
...... };

... values.forEach(function(value) {
...... accumulator["numberOfViews"] += value["numberOfViews"]
...... accumulator["totalTime"] += value["totalTime"]

...... for (var page in value["visitedPages"]) {
......... if (accumulator["visitedPages"][page] === undefined) {
............ accumulator["visitedPages"][page] = 0
......... }
......... accumulator["visitedPages"][page] += 1
...... }
... })

... return accumulator
}

Count of unique
elements?
> db.view.mapReduce(map, reduce,
{ "out": "userNavigationSurvey" })

# NOT AS WE WANTED

> db.userNavigationSurvey.find()
{ "_id" : "001", "value" : {
"visitedPages" : {
"example.com/001" : 1,
"example.com/002" : 1 },
"totalTime" : 6 } }

{ "_id" : "002", "value" : {
"visitedPages" : {
...

Count of unique
elements?
> var finalize = function(key, accumulator) {
... accumulator["averageTime"] =
...... accumulator["totalTime"] / accumulator["numberOfViews"]
... accumulator["numberOfUniquePages"] = 0
... for (var page in accumulator["visitedPages"]) {
...... accumulator["numberOfUniquePages"] += 1
... }
... delete accumulator["totalTime"]
... delete accumulator["numberOfViews"]
... delete accumulator["visitedPages"]
}

Count of unique
elements?
> db.view.mapReduce(map, reduce, {
... "finalize": finalize,
... "out": "userNavigationSurvey"
})

{ "_id" : "001", "value" : {
"averageTime" : 3, "numberOfUniquePages" : 2 } }

{ "_id" : "002", "value" : {
"averageTime" : 7.25, "numberOfUniquePages" : 3 } }

{ "_id" : "003", "value" : {
"averageTime" : 25.5, "numberOfUniquePages" : 1 } }

Count of unique
elements by steps
# STEP 1: CREATE THE BASE COLLECTION (WITHOUT UNIQUE ELEMENTS)

> var mapBase = function() {
... emit(this.user, {
...... "totalTime": this.time
... })
}

> var reduceBase = function(key, values) {
... };
... values.forEach(function(value) {
...... accumulator["numberOfViews"] += value["numberOfViews"]
...... accumulator["totalTime"] += value["totalTime"]
... })
}

Count of unique
elements by steps
> var finalizeBase = function(key, accumulator) {
... accumulator["numberOfUniquePages"] = 0
... accumulator["averageTime"] =
...... accumulator["totalTime"] / accumulator["numberOfViews"]
... delete accumulator["totalTime"]
... delete accumulator["numberOfViews"]
}

> db.view.mapReduce(mapBase, reduceBase, {
"finalize": finalizeBase,
"out": "userNavigationSurvey"
})

{ "_id" : "001", "value" : { "numberOfUniquePages" : 0, "averageTime" : 3 } }
{ "_id" : "002", "value" : { "numberOfUniquePages" : 0, "averageTime" : 7.25 } }

Count of unique
elements by steps
# STEP 2: CREATE THE COLLECTION OF UNIQUE ELEMENTS

> var mapUniquePages = function() {
... emit(this.user + "-" + this.page, {
...... "user": this.user,
...... "page": this.page
... })
}

> var reduceUniquePages = function(key, values) {
... return values[0]
}

> db.view.mapReduce(mapUniquePages, reduceUniquePages {
"out": "userUniquePages"
})

Count of unique
elements by steps
> db.userUniquePages.find()

{ "_id" : "001-example.com/001", "value" : {
"user" : "001", "page" : "example.com/001" } }






Count of unique
elements by steps
# STEP 3: UPDATE BASE COLLECTION WITH UNIQUE ELEMENTS COUNT

> db.userUniquePages.find().forEach(function(userUniquePage) {
db.userNavigationSurvey.update(
{ "_id": userUniquePage.value.user },
{ $inc: { "value.numberOfUniquePages": 1 } }
)
})

{ "_id" : "001", "value" : { "numberOfUniquePages" : 2, "averageTime" : 3 } }

Architecture

Webmachine

HTTP

HTTP

HTTP

Scalability

Webmachine

Nginx

Master/Master

Webmachine

Received by <account>
After <timestamp>?

function(document) {
if (document.from && document.to) {
var key = [ document.to, document.timestamp ]
var content = document._attachments["content"]
var outline = {
"id": document._id,
"from": document.from,
"timestamp": document.timestamp,
"type": content["content_type"],
"length": content["length"],
}
emit(key, outline)
}
}

After <timestamp>?

After <timestamp>?

> curl -X GET ".../mercurio/_design/message/_view/received_after"

{ "total_rows":3, "offset":0, "rows": [
{
"id": "ff35356344ee0e9928c212b52e36e6f3",
"key": [ "gabriele", 1263655442 ],
"value": {
"id": "ff35356344ee0e9928c212b52e36e6f3",
"from": "chiara",
"timestamp": 1263655442,
"type": "text/plain;charset=utf-8",
"length": 16
}
},
...
}

Results are
ordered by Key
{ "key": [ "chiara", 126 ],
"value": {
"id": "ff35356344ee0e992...",
"from": "gabriele" [ "chiara", 126 ]
}

== <
}

{ "key": [ "chiara", 128 ],
"value": {
"id": "0deff99666425bacc...", [ "chiara", 128 ]
"from": "gabriele"
}

<
}

{ "key": [ "gabriele", 120 ],
"value": {
"id": "9842063609746c661...", [ "gabriele", 120 ]
"from": "chiara"
}
}

Select with Key

{ "key": [ "chiara", 126 ],
"value": {
"id": "ff35356344ee0e992...",
received_after?
}
"from": "gabriele"
key=["chiara",126]
}

{ "key": [ "chiara", 128 ],
"value": {
"id": "0deff99666425bacc...",
"from": "gabriele"
}
}

{ "key": [ "gabriele", 120 ],
"value": {
"id": "9842063609746c661...",
"from": "chiara"
}
}

Select with range of Keys

{ "key": [ "chiara", 126 ], received_after?
"value": {
"id": "ff35356344ee0e992...",
"from": "gabriele"
startkey=["chiara",126]&
}
}
endkey=["gabriele",0]

{ "key": [ "chiara", 128 ],
"value": {
"id": "0deff99666425bacc...",
"from": "gabriele"
}
}

{ "key": [ "gabriele", 120 ],
"value": {
"id": "9842063609746c661...",
"from": "chiara"
}
}


{ "key": [ "chiara", 126 ],
"value": {
"id": "ff35356344ee0e992...",
"from": "gabriele"
}
}
[ "chiara", [] ]

{ "key": [ "chiara", 128 ],
"value": {
"id": "0deff99666425bacc...",
"from": "gabriele"
}
{ "key": [ "chiara", [] ],
}
"value": {
"id": "0deff99666425bacc...",
"from": "gabriele"
{ "key": [ "gabriele", 120 ], }
"value": { }
"id": "9842063609746c661...",
"from": "chiara"
}
}


"value": {
"id": "ff35356344ee0e992...",
"from": "gabriele"
}
}
endkey=["chiara",[]]

{ "key": [ "chiara", 128 ],
"value": {
"id": "0deff99666425bacc...",
"from": "gabriele"
}
}

{ "key": [ "gabriele", 120 ],
"value": {
"id": "9842063609746c661...",
"from": "chiara"
}
}

Received by “chiara”
After 126
"value": {
"id": "ff35356344ee0e992...",
"from": "gabriele"
}
}
endkey=["chiara",[]]

{ "key": [ "chiara", 128 ],
"value": {
"id": "0deff99666425bacc...",
"from": "gabriele"
}
}

{ "key": [ "gabriele", 120 ],
"value": {
"id": "9842063609746c661...",
"from": "chiara"
}
}

Push Received
Messages from Server

Check for Messages _changes?
received by <account.id> filter=message/received&
after <timestamp> by=<account.id>&
after=<timestamp>

Send Message Save Document
to <account.id> to: <account.id>

Push Received
Messages from Server

_changes?filter=message/received&by=<account.id>&after=<timestamp>

function(document, request) {
var receivedByMe =
document.to === request.query.by

var receivedAfterLastTime =
document.receivedAt >= request.query.after

return receivedByMe && receivedAfterLastTime
}

Backoffice as
Couch Application

CouchDB Vs MongoDB

More Related Content

CouchDB Vs MongoDB