Mongo DB for Java, Python and PHP Developers

MongoDB Introduction
for Java, Python and PHP Developers
Developers
for Java, Python and PHP Developers
Developers
Covers using MongoDB from a Java, PHP
and Python developer’s perspective. Uses
the Official MongoDB driver for Java,
Python and PHP as well as command line
tools for MongoDB to teach core concepts.

Copyright © Mammatus Technology Inc. | Licensed under Creative Commons.

Outline
• Theory and Architecture of MongoDB (49%)
• Setup instructions (9%)
• Code examples in JavaScript, PHP, Python and Java
(39%)
• 3% random

© 2012 10gen. MongoDB®, Mongo®, and the
leaf logo are registered trademarks of 10gen, Inc.

10gen in no way endorses this slide deck or Mammatus
Tehcnology Inc.

2

MongoDB
• NoSQL landscape full of contenders tackling big
data problems
• MongoDB very capable
• Document-oriented schema-less storage solution
• JSON-style documents to represent, query and
modify data
• Supports many clients/languages Python, PHP,
Java, Ruby, C++, etc.


Resources
• 10Gen site (http://www.mongodb.org)
– Great documentation and presentations
• InfoQ articles and presentations
• Wikipedia
• http://mongly.com/

4

Feedback welcome
• Send any and all feedback to
richardhightower@gmail.com

• Criticism welcome
– Prefer constructive criticism, but will take any and all
– Needed for continuous improvement of this slide deck and my
knowledge

5

Creative Commons
• This slide deck and all material therein are covered
under creative commons
• You can use all material in here as long as you don’t
copy it word for word and then use it for commercial
reasons
• Other material in here from other sources are
covered under fair use
• http://creativecommons.org/

6

Why MongoDB?
• MongoDB great active community
Supports:
• High availability,
• Journaling
• Replication,
• Sharding,
MongoDB commercial
• Indexing, http://www.10gen.com/what-is-mongodb

• Aggregation,
• Map/Reduce


MongoDB is a top job trend


Leader in the NoSQL space?

MongoDB seems to be
the clear mind-share leader

Cassandra a close second

9

Why do Developers pick Mongo?
• NoSQL, in general, can be more agile than full
RDBMS/ SQL
– problems with schema migration
– a lot of upfront design needed for RDBMS
– (or a lot schema migration later)
• MongoDB does not require a lot of ramp up time
– Easy to get started
– Many DevOps things come for free
– Easy on ramp for NoSQL
– Gateway drug?

10

Why PHP, Python and Java?

Might add Ruby later to the mix or just focus on these

11

Built for speed | Cache built in
• MongoDB was built with speed in mind
• Speed shaped architecture of MongoDB
• Uses binary protocol instead of HTTP text/
(CouchDB)
• Pads disk space around document
– faster updates
– uses more disk
• Uses memory-mapped files as default storage
engine, letting OS manage swapping
– Linux/Windows/Solaris really good at virtual memory...
MongoDB builds on top of this
12

Negatives of MongoDB
• Indexes are not as flexible as
Oracle/MySQL/Postgres or other NoSQL solutions
– Order of index matters, uses B-Trees, not very many options
like more mature solutions
• Realtime queries might not be as fast as
Oracle/MySQL and other NoSQL solutions
• Good enough if queries are simple
• Probably hits the sweet spot of 20/80 rule
• Not as mature as RDBMS
• Does not have full text search engine

13

Very useful still
• Every version seems to add more features
• Added journaling so they can have single server
durability
• Improved Replica’s with Replica Sets
• Replica Sets and Autosharding required very little
admin once running
• What it does, it does well...
– Can be combined with Relational database
– Can be combined with full text search (Solr)
– Can be combined with Hadoop

14

Who uses MongoDB?

Big names, big data
• MTV • SAP
• Craigslist • Forbes
• Disney • National Archives UK
• Shutterfly • Intuit
• foursqaure • github
• bit.ly • LexisNexis
• The New York Times • many more
• Barclay’s
• The Guardian

15

MongoDB Concept
• Oracle: Schema, Tables, Rows, Columns
• MySQL: Database, Tables, Rows, Columns
• MongoDB: Database, Collections, Document, Fields
• MySQL/Oracle: Indexes
• MongoDB: Indexes
• MySQL/Oracle: Stored Procedures
• MongoDB: Stored JavaScript
• Oracle/MySQL: Database Schema
• MongoDB: Schema free!


MongoDB Architecture

Copyright © Mammatus Technology Inc. | Licensed under Creative Commons. 17

Additional Mongo Features
• Geo Indexing: How close am I to X?
• File Storage
– Stores large files and file meta-data
• Capped Collection (like Ring Buffer)
– Older documents auto-deleted
• Aggregation
• Auto sharding
• Load sharing for reads
• High availability
• Speed or durability (journaling)

18

What do you need?
• Read scalability and high availability (HA)?
HA
– Use Replica Sets
• Write scalability?
– Use Autosharding (also just called sharding)
• HA, Read Scalability, and Write Scalability?
– Use Autosharding and Replica Sets
• You can start basic and add as your growth/needs
change
– Capacity planning, monitoring, determine needs

19

Durability
• Journaling added in 1.8, and is now default for 64 bit
OS for MongoDB 2.0
• Prior to that, you used replication to make sure copy
of operation was on replica
– MongoDB did not have single server durability, now it does
with addition of journaling
• General thought was/is durability is overvalued
• You can also force an fsync
• See links in notes

20

Replica Sets
• Drivers know the primary
• If primary down, Drivers Replica 1 Replica 2
know how to get new
primary
• Data is replicated after
writing
• Typical to have three in a
replica set Replica 0
• You can do more PRIMARY
• Load sharing for reads Read
Read/Write

Client Driver

21

Replica Sets Usage
• Business Continuity
• Data Redundancy
• High Availability
• Load sharing (reads)
• “Just works / NoOps (low ops)”

22

Replica Sets
• Non-blocking master/slave replication
• Auto failover
• Two or more nodes (usually three)
• No primary, master is nominated
• Share nothing architecture
• Brains in the client libraries
• Client libraries (Drivers) are Replica Set aware
• Client can block until data is replicated on all servers
(for important data)

23

Replica Sets
• You only write to the master, it then replicates to
slaves
• Replication is by default async (non-blocking)
• Slave data and write data can be out of sync
– There are workarounds
– You can force master to sync to master before
continuing (blocking, sync)
• Sync blocking is slower
• Async non-blocking is faster (eventual consistency)

24

Durability and Replica Sets
• Client libraries can do the following:

• Wait until write has happened on all replicas
• Wait until write is on two servers (primary and one
other)
• Wait until write has occurred on majority of replicas
• Wait until write operation has been written to
journal

25

Replica Sets
• Replica Sets and Autosharding go hand in hand for
mass scale out
• Replica Sets are good for failover and speeding up
reads, but...
• To speed up writes, you need autosharding

26

Sharding
• Sharding allows MongoDB to scale horizontally
• Sharding = partitioning
• Auto-shards
– load balances
– changes for data distribution
• Elastic adding of new nodes
• Supports automatic failover (along with replica sets)
• No single point of failure
• 90% of deployments don’t need sharding according
to Roger Bodamer
27

Non-sharded client connection
• Client Driver talks directly
to mongod process

28

Autosharded
• Three actors now:
mongod, mongos, and
Client Driver library
• Mongod is the process
• Mongos is a router, it
routes writes to correct
mongod instance
• Shares writing

29

Autoshard plus Replica Set
• Autosharding increases
writes, helps with scale
out
• Replica Sets are for high
availability
• There is a whole lesson on
sharding.

30

Sharding Topology
• Config Servers (mongod)
– contain versioned shard topology
– maps which shard has key
– used by mongos
– like DNS server for shards
• Mongos
– Shard router clients drivers talk to Mongos instead of mongod
directly
– Mongos uses Config Servers to find shard where key lives
– MongoD are shards that can be replicated

31

Large deployment

32

MapReduce
• Used for batch processing
• Similar to Hadoop,
• Massive aggregation possible through divide and
conquer
• Used instead of Group/By in SQL
– Also added simplified framework to MongoDB (aggregation
framework)
• Map and Reduce functions are written in JavaScript
– Executed on server, code next to data it is operating on
• Can copy results to results collections

33

MapReduce Theory

Image from http://code.google.com/p/mapreduce-framework/wiki/MapReduce

34

Incremental MapReduce
• Run MapReduce job over collections
• Run a second job but only over new documents in
collection
• Use reduce output to merge new data into existing
collection

35

Aggregation Framework
• Added in MongoDB 2.1
• Similar to SQL group by
• Before Aggregation framework, you had to use
MapReduce for things like SQL group by
• Easier to use than MapReduce

Table from 10 Gen
http://www.mongodb.org/display/DOCS/SQL+to+Aggregation+Framework+M
36

Aggregation Framework

Table from 10 Gen
http://www.mongodb.org/display/DOCS/SQL+to+Aggregation+Framework+M
37

MongoDB versus SQL

Compare contrast


SQL to Mongo

From http://www.mongodb.org/display/DOCS/SQL+to+Mongo+Mapping+Chart

39

Mongo versus SQL


40

Mongo versus SQL


41

Mongo vs. SQL


42

Getting Started with Mongo

Installing and using Mongo


Install (1 of 2)
• http://www.mongodb.org/downloads
• Extract
– ~/mongodb-platform-version/
• $ sudo mkdir /etc/mongodb/data
• Create file
– /etc/mongodb/mongodb.config
• $ cat mongodb.config
– dbpath=/etc/mongodb/data


Install (2 of 2)
• Link:
sudo ln -s ~/mongodb-platform-version/ /usr/local/mongodb
• Add to Path:
export PATH=$PATH:/usr/local/mongodb/bin
• Run the server:
mongod --config /etc/mongodb/mongodb.config


Run the client (type db.version())
$ mongo
MongoDB shell version: 2.0.4
connecting to: test
…
> db.version()
2.0.4
>


Client: mongo db.help()
> db.help()
DB methods:
db.addUser(username, password[, readOnly=false])
db.auth(username, password)
db.cloneDatabase(fromhost)
db.commandHelp(name) returns the help for the command
db.copyDatabase(fromdb, todb, fromhost)
db.createCollection(name, { size : ..., capped : ..., max : ... } )
db.currentOp() displays the current operation in the db
db.dropDatabase()
db.eval(func, args) run code server-side
db.getCollection(cname) same as db['cname'] or db.cname
db.getCollectionNames()


db.help()
db.getLastError() - just returns the err msg string db.printCollectionStats()
db.getLastErrorObj() - return full status object db.printReplicationInfo()
db.getMongo() get the server connection object db.printSlaveReplicationInfo()
db.printShardingStatus()
db.getMongo().setSlaveOk() allow this connection
to read from the nonmaster member of a replica db.removeUser(username)
pair db.repairDatabase()
db.resetError()
db.getName() db.runCommand(cmdObj) run a database command. if
cmdObj is a string, turns it into { cmdObj : 1 }
db.getPrevError()
db.getProfilingStatus() - returns if profiling is on db.serverStatus()
and slow threshold db.setProfilingLevel(level,<slowms>) 0=off 1=slow 2=all
db.getReplicationInfo() db.shutdownServer()
db.getSiblingDB(name) get the db at the same db.stats()
server as this one db.version() current version of the server
db.isMaster() check replica primary status db.getMongo().setSlaveOk() allow queries on a replication
db.killOp(opid) kills the current operation in the db slave server
db.listCommands() lists all the db commands db.fsyncLock() flush data to disk and lock server for
backups
db.logout() db.fsyncUnock() unlocks server following a db.fsyncLock()


Create Employee Collection
> use tutorial;
switched to db tutorial
> db.getCollectionNames();
[]
> db.employees.insert({name:'Rick Hightower',
gender:'m', gender:'m', phone:'520-555-1212', age:42});
Mon Apr 23 23:50:24 [FileAllocator] allocating new
datafile /etc/mongodb/data/tutorial.ns, filling with
zeroes…
..


Query DB (1 of 2)
> db.getCollectionNames();
[ "employees", "system.indexes" ]

> db.employees.find()
{ "_id" : ObjectId("4f964d3000b5874e7a163895"),
"name" : "Rick Hightower", "gender" : "m",
"phone" : "520-555-1212", "age" : 42 }


Query DB (2 of 2)
> db.employees.find({name:"Bob"})

> db.employees.find({name:"Rick Hightower"})
{ "_id" : ObjectId("4f964d3000b5874e7a163895"), "name" : "Rick Hightower", "gender" : "m",
"phone" : "520-555-1212", "age" : 42 }

> db.employees.find({age:{$lt:100}})
"phone" : "520-555-1212", "age" : 42 }

> db.employees.find({age:{$lt:100}})[0].name
Rick Hightower

> db.system.indexes.find()
{ "v" : 1, "key" : { "_id" : 1 }, "ns" : "tutorial.employees", "name" : "_id_" }

> db.employees.find({_id : ObjectId("4f964d3000b5874e7a163895")})
"phone" : "520-555-1212", "age" : 42 }

Java: Setup
• Download latest mongo driver
– https://github.com/mongodb/mongo-java-driver/
downloads

$ mkdir tools/mongodb/lib
$ cp mongo-2.7.3.jar tools/mongodb/lib

Create new Eclipse project in new Workspace


Java: Setup Eclipse (1 of 2)

• Right Click Project, Open Properties, Java Build Path->Libraries->Add Variable-
>Configure Variable


Java: Setup Eclipse (2 of 2)

• From Properties->Java Build Path->Libraries
• Click Add Variable, Select MONGO, Click Extend…, select jar
file you just downloaded


Java: Using Java Basics (1 of 2)

Out:
{ "_id" : { "$oid" : "4f964d3000b5874e7a163895"} , "name" : "Rick Hightower" , "gender" : "m" , "phone" : "520-555-1212" , "age" : 42.0}
{ "_id" : { "$oid" : "4f984cce72320612f8f432bb"} , "name" : "Diana Hightower" , "gender" : "f" , "phone" : "520-555-1212" , "age" : 30}

Using Java Basics (2 of 2)

Output:
Rick?
{ "_id" : { "$oid" : "4f964d3000b5874e7a163895"} , "name" : "Rick Hightower" , "gender" : "m" , "phone" : "520-555-1212" , "age" : 42.0}
Diana?
{ "_id" : { "$oid" : "4f984cae72329d0ecd8716c8"} , "name" : "Diana Hightower" , "gender" : "m" , "phone" : "520-555-1212" , "age" : 30}
Diana by object id?
{ "_id" : { "$oid" : "4f984cce72320612f8f432bb"} , "name" : "Diana Hightower" , "gender" : "f" , "phone" : "520-555-1212" , "age" : 30}

All of the above but in Python
• Install mongodb lib for Python
MAC OSX
$ sudo env ARCHFLAGS='-arch i386 -arch x86_64'
python -m easy_install pymongo
Linux
$ easy_install pymongo
or
$ pip install pymongo


Basic Python Operations (1 of 2)

Output:
{u'gender': u'm', u'age': 42.0, u'_id': ObjectId('4f964d3000b5874e7a163895'), u'name': u'Rick Hightower', u'phone':
u'520-555-1212'}
{u'gender': u'm', u'age': 30, u'_id': ObjectId('4f984cae72329d0ecd8716c8'), u'name': u'Diana Hightower', u'phone':
u'520-555-1212'}
{u'gender': u'm', u'age': 8, u'_id': ObjectId('4f9e111980cbd54eea000000'), u'name': u'Lucas Hightower', u'phone':
u'520-555-1212'}


Basic Python Operations (2 of 2)


All of the above but in PHP
• Install on PHP
$ sudo pecl install mongo

• Add to php.ini:
extension=mongo.so

• Restart apache
$ apachectl stop
$ apachectl start


Mongo PHP Basics (1 of 2)

Output:
array ( '_id' => MongoId::__set_state(array( '$id' => '4f964d3000b5874e7a163895', )), 'name' => 'Rick Hightower',
'gender' => 'm', 'phone' => '520-555-1212', 'age' => 42, )

array ( '_id' => MongoId::__set_state(array( '$id' => '4f984cae72329d0ecd8716c8', )), 'name' => 'Diana Hightower', 'gender' => ‘f',
'phone' => '520-555-1212', 'age' => 30, )

array ( '_id' => MongoId::__set_state(array( '$id' => '4f9e170580cbd54f27000000', )), 'gender' => 'm', 'age' => 8, 'name' => 'Lucas Hightower',
'phone' => '520-555-1212', )


Mongo PHP (2 of 2)

Output
Rick?
array ( '_id' => MongoId..., 'name' => 'Rick Hightower', 'gender' => 'm',
'phone' => '520-555-1212', 'age' => 42, )
Diana?
array ( '_id' => MongoId::..., 'name' => 'Diana Hightower', 'gender' => ‘f',
'phone' => '520-555-1212', 'age' => 30, )
Diana by id?
array ( '_id' => MongoId::..., 'name' => 'Diana Hightower', 'gender' => 'f',
'phone' => '520-555-1212', 'age' => 30, )


Shell commands in action
> show dbs
local (empty)
tutorial 0.203125GB

> show collections
employees
system.indexes

> show users

> show profile
db.system.profile is empty
Use db.setProfilingLevel(2) will enable profiling
..

> show logs
global

> show log global
Mon Apr 23 23:33:14 [initandlisten] MongoDB starting : pid=11773 port=27017 dbpath=/etc/mongodb/data 64-bit…
…
Mon Apr 23 23:33:14 [initandlisten] options: { config: "/etc/mongodb/mongodb.config", dbpath: "/etc/mongodb/data" }


Collection Methods (1 of 2)
> db.employees.help()
DBCollection help

db.employees.find().help() - show DBCursor help
db.employees.count()
db.employees.dataSize()
db.employees.distinct( key ) - eg. db.employees.distinct( 'x' )
db.employees.drop() drop the collection
db.employees.dropIndex(name)
db.employees.dropIndexes()
db.employees.ensureIndex(keypattern[,options]) - options is an object with these possible fields: name,
unique, dropDups
db.employees.reIndex()
db.employees.find([query],[fields]) - query is an optional query filter. fields is optional set of fields to return.
e.g. db.employees.find( {x:77} , {name:1, x:1} )
db.employees.find(...).count()
db.employees.find(...).limit(n)
db.employees.find(...).skip(n)
db.employees.find(...).sort(...)


Collection Methods (2 of 2)

db.employees.findOne([query]) db.employees.save(obj)
db.employees.findAndModify( { update : ... , db.employees.stats()
remove : bool [, query: {}, sort: {}, 'new': false] } ) db.employees.storageSize() - includes free
db.employees.getDB() get DB object associated space allocated to this collection
with collection db.employees.totalIndexSize() - size in
db.employees.getIndexes() bytes of all the indexes
db.employees.group( { key : ..., initial: ..., db.employees.totalSize() - storage allocated
reduce : ...[, cond: ...] } ) for all data and indexes
db.employees.mapReduce( mapFunction , db.employees.update(query, object[,
reduceFunction , <optional params> )
upsert_bool, multi_bool])
db.employees.remove(query)
db.employees.validate( <full> ) – SLOW
db.employees.renameCollection( newName ,
<dropTarget> ) renames the collection. db.employees.getShardVersion() - only for
db.employees.runCommand( name , <options> ) use with sharding
runs a db command with the given name where db.employees.getShardDistribution() -
the first param is the collection name prints statistics about data distribution in
the cluster


Basic shell commands (1 of 2)
> help
db.help() help on db methods
db.mycoll.help() help on collection methods
rs.help() help on replica set methods
help admin administrative help
help connect connecting to a db help
help keys key shortcuts
help misc misc things to know
help mr mapreduce


Basic shell commands (2 of 2)
> help
…
show dbs show database names
show collections show collections in current database
show users show users in current database
show profile show most recent system.profile entries time>= 1ms

show logs show the accessible logger names
show log [name] prints out the last segment of log in memory,
use <db_name> set current database

DBQuery.shellBatchSize = x set default number of items to display on shell
exit quit the mongo shell


Tips for scaling mongo
• Roger Bodamer
• http://www.infoq.com/presentations/Scaling-with-Mon
• Good ideas on EC2 shortcoming
• RAID configuration (RAID 10 for speed and scaling)
• Config Servers know where keys are, has key to
shard mapping, mongos refer to config servers

68

• Don’t use sharding unless needed
• 90% of deployments don’t need sharding
according to Roger Bodamer
– Are you Twitter, Linkedin, Facebook, Foursquare? No
– You probably not going to need it
• Replica Sets are more needed
– Why? HA, read scalability
• Mongos can live on primary box
• ConfigServer can live on a primary box
• http://www.infoq.com/presentations/Scaling-with-MongoDB

69

• Replicas should be on separate boxes

70

Backups
• You can use tools with mongodb
• Or, if your hardware supports shutdowns
• Sync to disk, shutdown cleanly, take a snapshot
• http://www.infoq.com/presentations/Scaling-with-MongoDB

71

Coming up
• Basic CRUD and queries: Slide deck showing a simple CRUD listing
website in Python, Java and PHP using MySQL and MongoDB
– Justing showing general operations
• Queries: Slide deck improving demo app to do more complex
queries
• Replica Set: Slide deck setting up simple Replica Set in Amazon
EC2 with boto scripts
– Durability modes with examples in Python, PHP and Java
• Sharding: Slide deck setting up Sharding in MongoDB
– Python, PHP and Java
• Map Reduce: Slide deck using sample app to do map reduce in
Java, PHP and Python

72

More to come
• This is an early version of this.... expect updates
• Looking for some feedback
• Want to grow it out, put it on github, etc.
• Do some compare and contrast between MongoDB,
MySQL, Cassandra, etc.
• Things you won’t get from a vendor
– complaining
– criticism
– shortcomings

73

About Author
• Rick Hightower, founder of ArcMind (RIP) and Mammatus,
father of five
• Former CTO of LearningPatterns and Trivera Technologies
(global training and consulting firms)
• Director of Development at 3 different places
• Author of five technical books
• Editor at InfoQ
• 20 years hacking code (C, Python, C++, Java, etc.)
• https://twitter.com/#!/RickHigh

74

About Mammatus, ArcMind
• ArcMind focused on Spring framework, JSF, Java EE,
Spring MVC, Grails, Groovy, Development, Tiles, etc.
– Started in 2003 ended in 2009
– Lot of clients (Boeing, Qualcomm, Bank of America, et.c), lots
of traveling
• Mammatus Tech (bumpy clouds) focuses on Cloud
computing, NoSQL, BigData, Map Reduce, Java, PHP,
Python, EC2, etc.
– Started in 2009
– http://cloud.mammatustech.com/
– http://nosql.mammatustech.com/
– http://www.mammatustech.com/

75

Mongo DB for Java, Python and PHP Developers

Recommended

More Related Content

What's hot (20)

Viewers also liked (12)

Similar to Mongo DB for Java, Python and PHP Developers (20)

More from Rick Hightower (19)

Recently uploaded (20)

Mongo DB for Java, Python and PHP Developers

Editor's Notes