Database Reference
In-Depth Information
Source Data
Source data coming into the data warehouse may be grouped into
four broad categories:
Production data.
Come from the various operational systems of the organization.
The significant characteristic of production data is its wide disparity.
Internal data.
These data relate to “private” spreadsheets, documents, customer
profiles, and sometimes even departmental databases. These are internal data, parts
of which could be useful in a data warehouse.
Archived data.
These are the historical data separated out from current data and
stored in archived media. Historical data are essential in a data warehouse.
External data.
This type of data emanates from outside the organization. Many
executives use performance indicators, competitor's market shares, financial para-
meters, and so on from outside sources.
Data Staging
This component of the data warehouse refers to the processes and
the area where preparation of the data takes place. In the data staging area, data
are cleansed, transformed, and prepared for loading.
Data Storage
This component includes all data storage in the overall data ware-
house, data marts, and other multidimensional databases. In the top-down method
of design, the overall data warehouse feeds dependent data marts. In a practical
approach, a collection of conformed data marts, by themselves, constitute the
organization's data warehouse.
Metadata
The metadata component in a data warehouse environment is similar
to the data dictionary or system catalog of a DBMS. But the metadata component
has more functions than a routine data dictionary. Metadata serves as a road map
for users to navigate within a data warehouse and examine the contents.
Information Delivery
This component includes a wide variety of methods for deliv-
ery of information to users. Users obtain information on-line by e-mail or through
the Internet or an intranet. Information delivery mechanisms include ad hoc reports,
complex queries, multidimensional analysis, statistical analysis, data mining appli-
cations, and executive information systems.
Management and Control
This component sits on top of all other components and
coordinates their services and functions within the data warehouse.
Dimensional Data Model
You are familiar with data modeling for operational
or OLTP systems. We adopt the entity-relationship (E-R) modeling technique to
create a data model. Operational systems possess the following characteristics:
•
Capture details of events or transactions
•
Focus on individual events
Search WWH ::
Custom Search