The data warehouse is distinctly different from the
operational data used and maintained by day-to-day operational systems. Data
warehousing is not simply an “access wrapper” for operational data, where data
is simply “dumped” into tables for direct access. Among the differences:
OPERATIONAL DATA
|
DW DATA
|
Application Oriented
|
Subject Oriented
|
Detailed
|
Summarized (or otherwise refined)
|
Accurate, as of the moment of access
|
Represents values over time, snapshots
|
Updated
|
Not Updated
|
Run repetitively and non-reflectively
|
Run heuristically
|
Requirements for processing not completely understood
before development
|
Requirements for processing understood before initial
development
|
Compatible with the Software Development Life Cycle
|
Completely different life cycle
|
Performance sensitive (immediate response required when
entering a transaction)
|
Performance relaxed (immediacy not required)
|
Accessed a unit at a time (limited number of data elements
for a single record)
|
Accessed a set at a time (many records of many data
elements)
|
Transaction driven
|
Analysis driven
|
High availability and not redundant
|
Relaxed availability and redundancy is a fact of life
|
Managed in its entirety
|
Managed by subsets
|
Static structure; variable contents
|
Flexible structure
|
Small amount of data used in a process
|
Large amount of data used in a process
|