BIG DATA MODULE 2 :

MONGODB

1
Summary

Chapter 1: Introduction to NoSQL

Chapter 2: Operation CRUD
Chapter 3: Requests
Chapter 4: Framework Aggregation

2
The Relational World
Characteristics:

Structured Data (Table/Schema)

Standardized Data (Normal Forms)

Standard (SQL)

Transactional (ACID)

Complex Query (Join)

Constraints(Data Integrity)
 NoSQL

NoSQL databases (understand Not Only SQL) responding to the

following issues :

- high availability
- high read and/or write performancethe
- processing of large volumes of data.
 NoSQL
A NoSQL engine is...

- A data structure

- An absence of constraints

- A modelling method

- A data schema... flexible

- A relationship with an object model of the querying

- Lack of transactions and compromises

NoSQL or SQL ?
CAP theorem
NoSQL solutions
NoSQL solutions

Key/value : in the manner of associative arrays of programming languages.

A key corresponds to a value.

Column oriented: a key corresponds to a set of columns, each with a value.

Document oriented: to a key corresponds sets fields/values that can be prioritized

Graph-oriented: data is modeled as nodes that have links between them.

NoSQL & Document Oriented: Example
MongoDB

MongoDB is a document-oriented database sponsored by 10gen.

A "document" is an entry in a database.

JSON that Mongo stores in binary (BSON)

MongoDB

● Document-oriented basis means:

that the stored objects are represented as a BSON document

Allows you to easily map to the objects we manipulate in our programs.

● MongoDB is schemaless.

No document schema is required to store the data.

MongoDB

MongoDB is a DBMS:

- document-oriented

- free

- scalable: replication, auto-sharding

- flexible: no data schema, full-text index

- written in C++.
Operational database
Document Data Model
Shell and Drivers
Vocabulary
Basic commands
Basic commands
CHAPTER 2

CRUD Operation
 Document JSON-Format
Official Site: [Link]

A json document is of the form: { } (object) or [ ] (array).

Document JSON-value
JSON-Example-1
JSON-Example-2
JSON-Example-3

Object with

nested arrays
and objects
Operation CRUD

The operations of CRUD MongoDB are available in the form of

functions /methods
(API) of a programming language and not as a separate language
(i.e. SQL).

Vocabulary:
MongoDB language
● Update

● [Link]()
● [Link]()
● [Link]()
● [Link]()

● Question

● [Link]()
● [Link]()
 Insert, Find & Count

[Link]({…})

The collection unique ID field is called “_id” and can be provided. If not provided an
ObjectID will be generated based on the time, machine, process-id and process
dependent counter.

“_id” does not have to be a scalar value – it can be a document, e.g. _id : {a:1, b:’ronald’}

[Link] || findOne ({…}, {field1 : true, …}).pretty()

//no argument will find all docs

[Link]({…})
Creating a document

Three ways to insert a document into mongoDB:

- Insert a document with insert

- Insert a document with update

- Insert a document with save

 Insert a document with insert

- A document is a table of key /values

-Documents is a list of table of key /values

- collection is the name of the collection you want add the document(s).

If the collection did not exist before, it is created (it is from this how collections are created)
Insert a document with insert
 Insert a document with save

Case: The document contain _id 🡺It replace

Case: The document does not contain _id 🡺 It makes a insertion.

 MAJ d’un document avec update

criteria is of the same form as for find

donnée_maj is a table of key / values defining operations on fields.

multi (Boolean):
- false (default): maj of an occurrence found (which one?)
- true: maj all occurrences

upsert (Boolean):
- false :update
- true : update or insert if not exists
Update a document with update
 Update an array with update

[Link]({ myQuery }, {myField: “newValue”, … })

🡺 replaces the existing document

[Link]({ myQuery }, {$set : {myField: “newValue”}})

🡺 Create or update myField

[Link]({ myQuery }, {$inc : {age: 1}})

[Link]({ myQuery }, {$unset : {myField: 1}})

[Link]({ myQuery }, {$set : {myField: “newValue”}}, {upsert:true})

🡺 Create or update document specified by { myQuery } with myField

 Arrays Update

.
 Delete a document with remove

- without any parameter : all documents are deleted (be careful, then)

- query is of the same form as for find, it refers to the documents that will be deleted.
Delete a document with remove
 Reading document with find

without parameters, all documents are returned

criteria is a table of key / values specifying operators on the fields of the documents sought

projection is a table to limit the fields you want to consult in the documents sought (this option will be
processed later)

Example:

[Link]()

[Link]( { ‘prenom’: ‘Camille’ } )

Reading document with find
 Reading document with findOne

The findOne function does the same thing as find but without bothering with a
cursor, when you want to retrieve a unique document (by its identifier,
for example).

If the query designates multiple documents, the first one found is returned.
Cursors
CHAPTER 3

Request
mongoimport
Operators
Comparison Operators
Operator on elements
 Operators Evaluated

Used to query the database

Operators: Logical
Query on a list (embedded)
Operator on a list
Query with notation « . »
Operator on a list
Aggregation framework

The aggregation framework extends the query and processing capabilities

of data.

With the new operators, developers will be able to sort and aggregate by
groups the data to which the queries relate and apply various operations to
them.
Pipeline
Example I
Example II
 Operations Pipeline

$match: query, equivalent to find

$project: Formats results, including removing/adding fields

$unwind: allows you to stream table, each element of the table will be treated as a
document

$group: Aggregates data

$sort: Sorting features

$limit: Limits the number of documents returned

$skip: excludes certain documents from the result

Cardinality
 $match

Use equivalent to find({...})

 $sort, $limit, $skip

$sort: Sorting documents

Memory-intensive; Can't use index (at least after grouping)

[Link]([ {$match: { state:"NY" } }, {$sort: { population:-1 } } ])

$limit: Limits the number of documents

$skip: excludes certain documents from the result

Makes only sense when you do a sort first
First skip – then limit (order of the stages in the pipeline matter)
 $project

$project: allows you to rework a document by adding, deleting, renaming … fields.

 $project

● Remove keys - If you don't mention a key, it is not included, except for _id,
which must be explicitly suppressed {$project: {_id: 0, …

● Add keys (also possible to create new subdocuments)

● Keep keys: {$project: {myKey: 1, …

● Rename keys / Use functions: $toUpper, $toLower, $add, $multiply

 $project

Another example
$unwind
$group
 Group operations

$sum: Add one to a key

mySum: {$sum:1}) to count number of elements

sum_prices:{$sum:”$price”}) : sum of variables

$avg, $min, $max: Average, Minimum or maximum value of a key

Create arrays: $push, $addToSet

categories: {$addToSet: “$category”}

Only useful after a sort: $first, $last

{$group:{_id:"$_id.state", population:{$first:"$population"}}}