Managing an Oracle Data Warehouse
© 2007-2016
by Burleson Corporation
* Understand the evolution
of decision support systems to data warehouse.
* Gain intimate knowledge of data - MYCIN.
* Employ parallel processors and terabyte storage.
* See examples of MPP data warehouses (Mead Data Central).
* Understand referential integrity and data cleansing.
* See past the marketing hype.
|
|
This two-day Oracle Data Warehouse training is designed to give the student a
firm conceptual understanding of the goals and techniques that are
used in a data warehouse. The Oracle Data Warehouse class explains the concepts and
techniques in plain English, and strips away all of the marketing
hype to give the student a clear understanding of how to construct
a data warehouse using the new technologies. The seminar begins
with a historical review of data warehousing and explains the basic
construction techniques for data warehousing and contrasts them
with traditional online teleprocessing systems.
The student will leave
this seminar with an understanding of data warehouse design
techniques for relational databases, including STAR schema design,
the pre-aggregation of data, and online analytical processing
(OLAP). In addition, the student will understand multidimensional
databases, and be able to understand how they are created and
maintained. This seminar also introduces data warehouse
"front-ends" for relational OLTP engines and explains how these
tools are used in real-world situations. Lastly, the student is
exposed to the data warehouse development life cycle where they
will understand data extraction, data cleansing, metadata
repositories, and practical tips for insuring a successful
warehouse project.
This course is designed for the working
Oracle professional and the amount of previous experience with
Oracle is incidental.
Previous experience with relational database management and SQL is
helpful, but this class is self-contained and has no formal
prerequisites.
|
This course was designed by Donald K.
Burleson, an acknowledged leader in Oracle database
administration. Burleson was chosen by Oracle Press to write
the authorized edition of Oracle High-Performance SQL tuning.
Burleson Corporation instructors offer
decades of real world DBA experience in Oracle features, and they
will share their Oracle secrets in this intense Oracle
data warehouse training.
Managing a Data
Warehouse
Don Burleson
Syllabus
Day One
9:00 - 10:00 The evolution of decision support systems to
data warehouse
DSS vs. Expert systems
The DSS evolution
intimate knowledge of data - MYCIN
support ad-hoc ("what if") queries
seamless access to distributed data
interactive in nature
either fully-structured or semi-structured queries
non-procedural queries ("show me more like this")
start at summarized level & drill-down into detail
GUI for visual presentation
IBM vs. Bill Inmon - evolution of the data warehouse
The new business environment
fast reaction to market changes
large volumes of non-fielded data
terabytes of legacy data - those who forget the past are
condemned to repeat it
the "invention" of OLAP
10:00-10:15 BREAK
10:15 - 11:00 Multi-dimensional databases - information
systems for decision support
pivot tables - three dimensional data (PC-BASED DEMO)
differing levels of summarization
11:00 - 11:30 Data mining - Conceptual overview
parallel processors and terabyte storage.
hands-off analysis - identification of unobtrusive trends within
data
Intelligent agents - used to identify trends based on "hints"
from users.
linear regression example - trumpet of doom
11:30 - 1:00 LUNCH
1:00 - 2:00 Hardware & Software advances with data
warehousing
Parallel processing hardware
MPP& SMP architectures
Examples of MPP data warehouses (Mead Data Central)
Software advances
AI-based query engines (ConQuest, Folio, Fulcrum)
Parallel database software
multi-dimensional databases (Essbase)
Enterprise computing architectures
common query languages - SQL wrappers UniFace EDA/SQL
data warehouse tools - Prism, Vality
2:00 - 2:15 Break
2:15 - 3:30 Introduction to data warehouse design
Data collection analysis
STAR schema design
de-normalization - pre-joining tables
The Role of Metadata
Why is metadata important? (changed procedures, usage)
What are the current metadata tools?
How to choose/create a metadata repository
3:30-3:45 BREAK
3:45-4:30 Pre-aggregating relational data for relational
warehouses
EXERCISE
Day Two
9:00-9:45 Exercise review
9:45-10:30 Populating the data warehouse:
Data-driven vs. application-driven design
Data is collected into a data-driven design
Keeping the warehouse current
Asynchronous updating strategies
Data extraction and cleansing
Extracting from the warehouse Slicing the data warehouse -
operational
"views" across functional areas vs. functional slices.
Process mgt. for staging, population, and slicing data
Referential integrity and data cleansing
Enforcing business rules across diverse systems
Sizing the Data Warehouse
Estimation of initial sizes
Relational determination
Sizing for non-relational inputs
Forecasting future size requirements
Rolling/archiving of obsolete data
10:30-10:45 BREAK
10:45-11:30 Data Warehouse development life cycle
The feasibility study - cost/benefit analysis
the issue of intangibles
The black-hole argument
The pilot study
survey existing applications
review external data sources
Marketing the project
packaging the warehouse
training on data warehouse usage
maintenance
Implementation Hints
New Trends - Data Marts, Data Trolling
Evaluating warehouse tools & products
Seeing past the marketing hype
Key elements to look for in a product
Market share
User success stories
Robust functionality/extensibility
11:30-1:00 LUNCH
2:30-5:00 Site-specific data warehouse issues
Open discussion - Bring your questions/issues
|
