4. OLTP vs. OLAP
OLTP OLAP
Stands for Online Transaction
Processing (Known as :
Operational systems)
Online Analytical Processing
(Known as: Data warehouse)
Source of data Operational data; OLTPs are
the original source of the data.
Consolidation data; OLAP data
comes from the various OLTP
Systems
Purpose of data To control and run
fundamental business tasks
To help with planning, problem
solving, decision support
Types of queries Relatively standardized and
simple queries Returning
relatively few records
Often complex queries involving
aggregations
Type of data Daily transaction data. Mostly
doesn't keep history
Historical data
5. OLAP Contd..
â˘OLAP is also known as Data warehousing
â˘Definition:
â˘subject-oriented,
â˘integrated,
â˘time-variant,
â˘non-volatile collection of data in support of management's decision
making process.
6. Why Data ware housing ?
Why Data
warehousing?
Who are my
Customers and What
product they
frequently buy?
Which Customers
are most likely to
go for
Competition?
Who are my
Customers and What
product they frequently
buy?
What Products
Promotions have
biggest impact on
Revenue?
What is the most
effective distribution
channel?
7. Data warehouse Architecture
Enterprise
Data
Warehouse
Sales
(Data Mart)
HR
Reporting
Tools
OLAP Tools
Ad Hoc
Query Tools
Data Warehouse
Database
lOracle
lSQL Server
lTeradata
lDB2
Data and Metadata
Repository Layer
ETL Tools:
lInformatica Power center
lPentaho
lAbinitio
lOracle Warehouse Builder
lCustom programs
lSQL scripts
Extract,
Transformation,
and Load (ETL)
Layer
l Cleanse Data
l Filter Records
l Standardize Values
l Decode Values
l Apply Business Rules
l Dedupe Records
l Merge Records
ETL Layer
Operational/OLTP
systems
l PeopleSoft
l SAP
l Siebel
l Oracle Applications
Products
Sample Technologies:
Execution
Systems
l CRM
l ERP
l e-Commerce
External
Data
l Purchased
Market Data
l Spreadsheets
Source system Presentation
Layer
Reporting
l Cognos Reports
l Business Objects
l MicroStrategy
lData Mining Tools
l Portals
lCustom Tools
lHTML Reports
8. ETL
ETL Stands for Extract â Transform â Load
Extract >Extract data from various source systems as efficiently as possible
Transform >Clean and perform calculations on extracted data
Load >Load data into target Data Warehouse (ex: Oracle, MSSQL databases)
ETL can be implemented using:
â˘By scripts(Shell, Perl, Python)
â˘PL/SQL
â˘Transformations: Java and C++
â˘Using ETL Tools (E.g. Informatica, Abinitio, Pentaho etc.)
9. Benefits of Informatica
â˘Process different types of Data sources
â˘Initiate ETL Projects Quickly and Effectively
â˘Automate most ETL processes for fewer errors and greater productivity
â˘Visual flow and self-documentation