Databases Reference
In-Depth Information
Chapter 42
Migrating Data to an
Integrated Database
James A. Larson and Carol L. Larson
RDBMS
.
However, migrating data
into a unified relational database can be quite challenging. To perform a
successful migration, database administrators need to know how to:
ARE
A
GREAT
BOON
TO
DATA
ACCESS
• Detect and resolve structural inconsistencies.
• Detect and resolve value inconsistencies.
• Balance the tradeoff between the cost of maintaining an accurate data-
base with the cost of mistakes due to inaccurate data.
STRUCTURAL INCONSISTENCIES
Inconsistencies among data files can be classified as either structural or
value. Structural inconsistencies are incompatibilities among the schemas
or file descriptions and should be identified and resolved before migrating
data between the source and target files. Value inconsistencies are differ-
ences in values of corresponding schema items and can be identified only
as data is moved and integrated from multiple sources into a single target
file. Database administrators deal with each of these two inconsistencies in
very different ways.
Structural inconsistency is the situation where the organization and for-
mats of the target schema are different from the organization and formats
of the source schema. Types of structural inconsistencies include naming
issues, different abstraction levels, inconsistent coding structures, incon-
sistent data formats, and unneeded or missing data.
Naming Issues
The first task facing database administrators in determining how the
data items of a source file relate to the data items of a target file is the iden-
tification of corresponding data elements. Database administrators often
use data element names as clues to map data elements of a source file into
the data elements of a target file. Exhibit 1 illustrates the notation used to
Search WWH ::
Custom Search