Dp 900 AZURE FUNDAMENTALS

Core Topics
Basic
Terminology
Core Data Concepts 15 20

Relational Data 25 35

Non Relational Data 25 309

Data Analytics 25 30
 BASIC

TERMINOLOGY
Cloud
computing
Types of cloud
Scaling Types
Introduction to Azure

Azure Service's
CLOUD Computing means delivering
computing services over the internet

Cloud Benefits
scalability
processing
storage
tibierance
Cost Effective

Security

Types Of CLOUD
public Cloud Aws AZURE Gcp
private cloud I'm cloud Vmware
one organisation
Hybrid Cloud min of 15,9
p EtÉe
SCALING TYPES

Vertical monolythic
1 Node Architecture
Horizontal Distributed
Architecture
 AZURE is
microsoft's cloud computing
platform which provides services

SERVICES
Compute UM's
Storage Data lake Blob
Storage
Databases SQL DB Cosmos DB NosalDis

Networking Vnet Load Balancer

Data Al Data
Factory Synapse
Databricks
DATA LAKE vs DW vs DB
Analytics Transactions
missy

semistructured

Read

Note Azure DW primarily handles structured

Relational but can semi structured Data
Query
Semi Structured Data is
usually stored in Data Lake
accessed using external tables or polybase

BLOB
Binary Large Obtect
CORE
DATA
CONCEPTS
Represent Data
Features of Data variety
options for Data storage
Formats for Data files
Types of Databases
Common Data workloads
Transactional OCTP
Analytical OLAP
Roles and Responsibilities
Types OF ANALYTICS

Descriptive what happened

summary of existing Data
example Today's sales

Diagnostic why it happened

Deep diving to understand cause

example sales by State Gender

happens in future
predictive what

Based on
past Trends

example weather prediction

Prescriptive what to do

Advice on best approach

example movie Recommendations
based on likes of user

Cognitive AI ML
Learn Improve over time

example Self Driving Cars

DATA Raw facts Images etc

Data at Rest Blob

continuously flowing
Types Of Data variety

searchable

asn.it i i n nEEn.ns

FEATURES OF DATA Variety

CHOOSE RIGHT STORAGE

Key factors Search Perform Cost

AZURE SQL DB
AZURE ULTRA DISK
AZURE BLOB
ADLS Synapse
 Relational workloads
OLTP vs
OLAP
we have Transactional Analytical

Olyppatabase pupparehouse
Day to Day Reports
operations Deep Analytics
Azure SQL DB Azure Synapse
Cosmo DB AAS

simple Historical
OLTP
Database

OLAP
Data
warehouses

On when to use Data warehouse

when queries are long complex Analytics
Data from multiple sources structured semi
when Data need further ETC or ELT

ETL vs ELT
Common DATA ROLES
Responsibilities
Database Admin DBA Guardian of DB's
Manager Databases
Availability performance
Assign permissions
Manage Data backups
Handles Restores
Disaster Recovery

Data Engineer Build pipelines to ensure enough

data is there
for Analytics AI
Data Migration without an Error
Data Cleaning Routines
Apply Data Governance privacy GDPR policies
Export Data
Import pipelines

Data Analyst Analyze the Data

Create value from Data
EDA
Reports visualize

Other Roles
Data Scientist
Data Architect
Application Dev
Software Engineer
RELATIONAL

STORAGE
Features
Normalization
Azure Services
Azure SQL Database
Database Security
Query Tools
 RELATIONAL DATABASE stores data in
form of
Tables which are related to each other using
common Column based on
primary foreign key

1
primary key Unique No Null
2
Foreign key contains duplicates but references
to
3
Composite
primary key
key combination of 2 or more
keys
to form a
primary key
3 Index allows for efficient retrieval of Data
in SQL Queries

primary key is Default Index

we can define other indexes
example put an index on Last Name

4 View are quest virtual table based on

other Real tables

they do not store data
Any changes in view won't change main table
No
privacy access to
underlying Tables
NORMALIZATION is the process of breaking
down Large tables into small one's to
avoid Data
Redundancy
maintain Data integrity
to avoid insert update Delete anomalies

Norma Forms INF 2nF 3nF BCNF etc

INF Atomic values

example 123 123 124

2nF No partial Dependency

example product ID OrderID products
primary key composite
here productsonly dependent on
are

half of primary key partial Dependence

Solution Remove any data that is not
related to entire
primary key
3nF No Transitive
Dependency
All Data field must be related to P.key
of
Manager ID Employee ID Pkey
E manager partner Name Employee name

Manager Son's Name Job Title

Manager's partner son is not dependent on Emp ID

Solution Remove data that is not related to p key
any
SQL case insentive Microsoft TSQL
is used to interact with databases

enabled by SQL

SQL COMMANDS

DBA

Rename

CREATE VIEW
Create View myview as
SELECT Customer ID Sales FROM dbo sales

CREATE STORED PROCEDURE

Create Jarameter

CREATE INDEX efficient Data Retreival

Note once inden created on

city rather checking Kashmir

in all rows it will have
inden set for it which makes
data retreival Quick
we wont create inden for all columns
because it also takes
computation power
 AZURE SERVICES

Vin's BLOB

To Deploy Azure

365 power BI

AZURE RELATIONAL DB's

Azure SQL DB SQL Server in VM
5GB 4TB Storage
2 80 Cores
5 month

Single DB Allocate resources to specific DB

Elastic DB Allocate resources to Group of DB's pool
Multiple DB's in Server
ADVANTAGES
mostly compatible with SQL Server
scale
You can
easily
AZURE SYNAPSE ANALYTICS SQL Spark
More than a Data it
warehouse as
helps
with Data warehousing and Big data
Analytics as well

Data warehouse Data Analytics

SQL Spark
CREATE AZURE SQL DB

BE.IE 75 iththmDev
production 5001m
Vcore plan which let's us choose
the number of cores
Storage Redundancy

GIreaiim.to s'inIiipie locations saved

Defender for SQL 20 month

Sample Data Adventure works
 ARM AZURE RESOURCE MANAGER MODEL
Azure uses son for data formats in Azure SQL DB
ARM is used to automate deploying Resources
How does it work
Resource Group Automation ARM Templates
Finally we can
Deploy it again or we can use
templates to run it as script change parameters

DATABASE VS SERVER

SQL DATABASE SECURITY

1 access
Networking public private
we Rules
can customize access based on Firewall
Rule Name Start Ip address End IP address

2 Microsoft Entra Active Directory

or
identity
access Security IAM
ssco
anam.com
Secure
Role Based
password less
3 Microsoft Defender is a
security solution
that protects database from Cyber threats
Threat Detection If someone tries to download all data

4 Data Encryption storing data in Unreadable

format even if data is stolen it makes sure

no one read it
can Databases by defant are

encrypted in Azure but we can turn it off

5 Logging Auditing setting up alerts

RELATIONAL QUERY TOOLS

QUERY EDITOR

QueryEnnor

AZURE DATA STUDIO SSMS 09

light weight FULL SQL Server Admin
modern DB management
Cross platform

Form

SQL SQL SERVER TSQL

ORACLE POSTGRE PLSQL procedurallanguage
The query Syntan won't match
 NON
RELATIONAL

STORAGE
Create Storage Account

Storage Types
Non Relational Databases
Now Relational Data types
Cosmos DB

Create Cosmos DB Account

Features of Cosmos DB
CREATE STORAGE ACCOUNT

Faster performance
SSD
for blobs

Redundancy
LocalLRS 3copies Datacenter
Zone ZRs 3copies 3 Data
centers
I Geo GRS 6 copies
3 in primary Region
3 in Remote Region

Access HOT Freauenty Used Data

Tiers
COLD Rarely used Backups cheap

Gen 2 STORAGE

Big Data Data lake

STORAGE TYPES

For Blobs unstructured

For File sharing

Message Queues
Schema less Tables
CREATE CONTAINER BLOB STORAGE unstructured Data

we can change
the storage
access level
container inside
Not proffered
container

upload Any Type Data inside the container

Change Access Tier Hot cool etc

VIEW EDIT
Mar 2 MB Blob file can be viewed
ALLOW ACCESS TO CONTAINER FICA

Storage Account Configuration

Container File Change Access

Level

Access In PYTHON
FILE SHARES is a cloud based file storage
service that allows you to store share access
files guest like a network drive It works like
a shared
folder that multiple users or application
can access from anywhere
Man Limit 100 Tb
Connect windows Mac Linux

Copy Script and use in Cmd Line

Connect

AZURE TABLE STORAGE Semi Structured Data

is not like SQL Database It is counted as

Non Relational Storage No Schema

Add Table Storage Browser Add Entity

No Relational Element Dynamic Schema

1 Add
Table

2
NON RELATIONAL DATABASE does not used fired
schema or relational tables NO SQL Gtructure
Highly scalable Big Data XML json
Dynamic Schema
Documents key value column
family Graph
NON RELATIONAL DATA TYPES

DOCUMENT COLUMN FAMILY

C DB core SQL cosmos DB Cassandra API

KEY VALUE GRAPH OTHERS
FB Time series
CSV Json in
Blob
Azure Search

Cosmos DB Table Apl Graph API

CHOOSE NON Relational DB
tomillisecond
latency
Filice a

DataBackup

DataAnalytics

NON RELATIONAL DBS IN AZURE

Important Concept
 Azure Cosmos DB is a no SQL

database fully managed by Microsoft

designed for large data multimodal
Low 10ms
latency Ap
high availability
CREATE Cosmos DB Flexible database to

handle of Data
large amounts

primary DB Relational Azure SQL

Non Relational Cosmos DB

Most common
Json

NOSÉE
Ap

MULTIPLE MODEL POSTGE Ap Only Relational

soooooo

Can I save images in Cosmos DB

 Automatic Scaling

payasyougo

Multi region writes Data can be

written from multiple region
nine
inatage

Query Cosmos DB NoSQL

Overview

Create container
 Partition key
If data gets
we can use
large
our
obit to partition
date based on
a column like
Country etc
panicat
country

ADD ITEMS
so addnew Row
0
L view
manual

Auto

Is SQL inside Cosmos DB case Insensitive

ARM TEMPLATES TO MANAGE COSMOS

DB
Automation Export Template
we often ARM so that we
use can
save these codes in Github because if
any change is to be made we can

directly make those in code it is

also called IAC Infrastructure as code
Cosmos DB
Security

TÉÉ promised regenerate

Cosmos DB GEO REPLICATION

It is alot more easier to replicate
data globally we can
provide
head read write access but it
2x's or 3ns the cost

choose

line
gemont
 DATA

ANALYTICS
Data workloads
Data warehouse
Azure synapse
Microsoft Fabric
ADF
pipelines Triggers
Visualization
power BI
Types of Reports
components
Common DATA WORKLOADS

simple Historical

integrity isthere

forpostgre SSAS

General use complen use

write Heavy Read Heavy
Constantyupdates No updates
A2
125 1 D synapse

DATA WAREHOUSE
is a
large database where data
is from different sources is collected

cleaned and organized so that

companies
can analyze it and make better decisions
or
Giant library of data where
is sorted everything
ready for reporting analytics
o
Stores historical Data
OLAP complen queries Report
Combines data from multi systems
 when to use Data warehouse
When queries are
long running complen
When Data Needs further cleaning ELT or ETL
Archiving whenhistorical data needs to be moved
from Day To Day Systems
When we need to integrate data from
several sources

Azure Synapse Dwt more

can store both structured

Semi structured
Unlike Traditional Data warehouses

Components OF MODERN DATA WAREHOUSE

ELT
All Data comes
cubeDimensions
fromothersources 11 1
Queries

Hiiiii
ETHELT
HEYL
inÉiigence
orchestration
service

iii IIII
fiI I i i Eammin

Data from multiple sources ADF

Unstructured Data Data lake Blobs Databricks
Structured Data Synapse DW B
power
MICROSOFT FABRIC SaaS tool which

brings Data factory One lake

synapse
power BI all under one
roof No more

switching between Synapse Data etc

Factory

Data governance

BEN FITS Unified Saas model

End to End pipelines
Data Governance
security
Ingestion Transformation Analytics
Key FEATURES
One Lake A unified data lake that serves
as a
single storage layer onedrive

Synapse powered Engine combines capabilities

from Azure Synapse Analytics for powerful
Querying Data warehousing

power BI Data Science

Microsoft purview control center for data

security compliance
How was it before Fabric
orchestration

Before my P
After One Integrated
software

DATA ENGINEER RESPONSIBILITIES

AZURE DATA FACTORY is a data
Orchestration and integration service used to
build ETL ELT pipelines for moving
transforming data across hybrid cloud systems
supporting both Real time Batch processing
OR
It is a
pipeline buider that helps you
collect clean and deliver data
automatically
WORKING
Brings Data from external sources
Data Transformation ETL ECT
Scheduled Jobs Automate Transformations
What is Orchestration in Azure Data
Factory
In ADF it to refers to controlling how data
flows from sources to destination tasks like
1 Extract Data from sources SQL Blob
2 Data
Transform clean filter 70in
3 Load into Destination Azure Synapse Data Lake
4 Trigger workflows or schedule based on events
5 Handle failures automatically
CLOUD VERSION OF 5575
pipelines
DATA FACTORY Pipelines amplifier
is a series Data related tasks activities
of
where to get the Data Extract
what to do with Data Transform
where to send Data Load
PIPELINE TRIGGER
Pipeline Run pipeline Execution

manuailly Trigger
Trigger the An event that causes pipeline
pipeline to run

Main Types

Every Day 8 00pm

Every hour
we can set up start End

Can be set to run in the past

Good when pipeline is time period
specific

8 00 9 00
9 00 10 00
o overlap
VISUALIZATION mostly focus on
power BI

Types OF REPORTS

Paginated Interactive Dashboard

paginated Reports Designed to live on page

power BI Report Buider Download separately

Finally publish to service

Interactive Report Designed to be viewed on

screen interacted with
visuals slicers etc
add remove column
sort etc

Dashboards Summarized view of our Reports

POWER BI WORKFLOW
Load Data
Transform power Query
Data modelling
Dan measure calculated columns
visualize
Analyze
publish
 EXAM QUESTIONS

Q what are two purchasing models in SQL Database

DTUs core

Qs If you set up SQL DB with no access which

type
of user can connect to it
No one

Q compatibility with SQL Server in own Environment

your
SQL Server in VM 100
SQL managed Instance I 100 close to 100

Q1 Adding Database effect on cost

Single Database plan Yes

Elastic pool No we can have upto 100 DBS
Q which of the following Azure services use the
SQL Server Database engine
Azure SQL Db
SQL server in VM Not MySQL
SQL managed Instance

Q1 Single Table Storage 5 PB

Q1 Does scaling affect applications using DB
No it wont affect that time
01 If SQL Database is setup public endpoint
which type of user can connect to Database
No user because