MongoDB Introduction, Installation and MongoDB Schema Design: How to Think Non-Relational

Key Points for Basic context of MongoDB(stoped at 30 minis or so):

The official site is here: http://www.mongodb.org/ and there is a web-based command line interface for trying the basic operations;

After download and unzip the gz file, then mkdir /mongodb/data;

To start the mongo db: ./bin/mongod --dbpath /mongodb/data

to connect the mongdb ./mongo which will connect to the localhost and test database by default;

To save a demo JSON data like this :       db.mycol.save{"name":"sch", "interests" : "mongdb" };

To retrive the demo data like this :            db.mycol.find() which no need to do any joins bewteeen tables;

To create a B-Tree index on interests:       db.mycol.ensureIndex({interests:1})  then db.mycol.find({Interest: Mongodb});

To explain the query plan:                        db.mycol.find({interest: Mongodb}).explain() ;

To created sparse index:                          db.shapes.ensureIndex({radius:1}, {sparse:true}) which will tell database not  to create index for the data that does not have this  field                                                        which is quite space saving and performance efficent

The 'dot' operator:                                    db.books.ensureIndex({comments.author:1}) //create index on mnested documents

                                                              db.books.find({comments.author: "sche"})

                                                              db.books.ensureIndex({comments.votes: 1})

                                                              db.books.ensureIndex({commnets.votes: {$gt: 50}}) //find all books iwht more than 50 votes 

 

In the above DB shell cmd. mycol does not need predefined in the MongoDB and it will automatically created by database;

There is a web site for mongodb which is:   localhost:28017;

To create an index on the "tags" array which it is called multi-key indexes;

Key Points for MongoDB Scheme

Ways to model data: normalize or denormalize

Comparision of RDBM and MONGODBRDBMSMonggoDBTable           CollectionRowsJSON documentIndexIndexJoinEmbedding and Linking

Working with documents

Evolving a schema

Queries and indexed

Rich documents

compared with RDBMS(by contrast):  No need to access data from different locations and memory and tables,

                                     No need to worry about the wrong query plan which could impact the performance heavily

MongoDB is a JSON database and MongoDB shell is a javascript driven interacive interface;what is the BSON?

How can we manipulate the data?

Dynamic Queries

Secondary Indexes

Atomimc Updates

Map Reduce

No joins for MongoDb because it will make scalability horizontal impossible instead of only scalaer up(not out)

Query operators:

$ne,   $in,       $nin,    $mod,$all

$size, $exists, $tyep,  $lte

Conditional operators:

$set,        $inc,   $push,  $pop

$pushAll,  $pullAll

            Extending the Schema:

new_comment = {

author:   "schen",

data:      new Date(),

text:       "great book",

votes:     5

}

 

db.books.update(

{ text:"moon"},

{ '$push' : {comments:new_comment},    //for array purpose

   '$inc'   : {comments_count: 1}

}

)

One to Many Model

Embeded array/array keys

-$slice operator to return subset of array

-some queries hard

Embeded tree

-single docuement

-natural

-hard to query

normalized(2 collections)

-most flexible

-more queries

Referenceing vs. Embedding

-Embed when the 'many' objects always appear with their parents

-Refence when you need more felxibility

Many to Many Model