2. Contents
Introduction and writers intro
Normalization and De-normalization
Methodology
Single Dimensional Queries
Multi Dimensional Quires
Conclusion
3. Introduction to Topic
Two most common processes in database design
are data Normalization and De-normalization, this
research discuss opimization using Hierarchical
Denormalization
Mortez Zaker is a research student in Advance
Databases and data warehouse, he is System
Analyst and specialized in database.
Somnuk Phon-Amnuaisuk done B.Eng and have done
Ph.D in Artificial Intelligence.
Dr.Su-Cheng Haw’s has research in XML Database.
Her favorite fields Data Modeling, Design, Data
Management, Data Sestamatic etc.
4. Introduction
Datawarehouse is basically foundation for Decision
Spport Systems, which involve huge collection of
information, extractions of data from different sources
which is available for On-line Analytical processing
(OLAP)
DW involve number of join operations, incurring
computational overhead, multi-dimensional grouping and
aggregation Operations
Due to nature of DW Performance and efficiency are
most required qualities.
Researcher compared the results of Hierarchical
De-normalization and how much it effect
performance.
5. Normalization
1. Normalization
Normalization is technique in which we
120 reformat and design database and break in in
such way that it can give us more speed and
integrity constraints and restriction of business
rules.
70
2. De-normalization
50
Database is transferred from Normalized to
De-normalized to create data ware house.
30
2003 2004 2005 2006 Data Warehouse
6. Pakistan
Hierachy
Each member in a hierarchy is
Known as a “node”. The top node is
called root
And bottom nodes are leafs notes.
Punjab A parent
Frontier
Node is a node that has children
And a child node is a node which
Belongs to a
Lahore
A parent
Peshwar Multan
Mardan
Qasa
Gulberg Defence
Khawani
Street
7. Methodology
Researcher used to compare efficiency of De-Normilzation and normalization
processes and analysis the performance of data models, used series of queries on
some column for evaluation.
Normalized data set
De-Normalized data set
8. Query Set
A set of query Benchmark has been Query
Selection used for frequent query application like
Star-Schema in the data warehouse
Queries designed on based business
Business analysis missions.
Evaluation was performed by using
Evaluation calculating time using all queries in Record Results
benchmark
11. Conclusion
Researcher used their methodology and research to know exactly what is effect of
using hierarchical de-normalization. The come to following conclusion
The findings confirm that most probably hierarchical denormalization have the capability
of improving query perfromance since they can reduce the query response times when
the data structure in Data Warehouse is engaged in several joins operations. The result
can help researcher in the future to develop general guidelines which can be applicable to
majority of database designs.