Information Technology Reference
In-Depth Information
6FKHPD$
6FKHPD%
7UDQVIRUPHG6FKHPD$
&ODVV
[
&ODVV
[
&ODVV
[
&ODVV
[
&ODVV
[
&ODVV
[
$WWULEXWH
$WWU
$WWU
$WWULEXWH
$WWU
$WWU
$WWULEXWH
$WWU
$WWU
$WWULEXWH
$WWU
$WWU
$WWULEXWH
$WWU
$WWU
$WWULEXWH
$WWU
$WWU
!
6FKHPD$
6FKHPD%
7UDQVIRUPHG6FKHPD$
&ODVV
[
&ODVV
[
&ODVV
[
&ODVV
[
&ODVV
[
&ODVV
[
$WWULEXWH
$WWU
$WWU
$WWULEXWH
$WWU
$WWU
$WWULEXWH
$WWU
$WWU
$WWULEXWH
$WWU
$WWU
$WWULEXWH
$WWU
$WWU
$WWULEXWH
$WWU
$WWU
!
Fig. 7.17
Merge classes by associations
always overrides total participation in the integrated schema. It is because total
participation is a subset of partial participation.
When dealing with data type and subtype conflicts, the association/relationship
relation is used for resolution. To illustrate this, assume we have an attribute
De-
partment
of the entity
School
in one schema and an entity
Department
in another
schema. To resolve the data type conflict, a 1:n relationship is formed in the inte-
grated schema to link up these two entities.
Step 2. Merge classes and relationship into frame model metadata
Input: Existing schemas A and B
Output: Merged (integrated) schema X with semantic constraints captured by
frame model metadata
Classes are merged using the union operator if their domain is the same. Other-
wise, abstractions are used under careful user supervision. By examining the same
keys with the same class name in different database schemas, we can merge the enti-
ties by union. The integrated class takes all the attributes from both entities. Abstrac-
tions like generalization and aggregation are used in merging classes in different
input schemas when they fulfill the semantic condition. The details are as follows.
Substep 2.1. Merge associations by capturing cardinality.
The integration can be based on the richer data semantics of 1:n association and
which can be specified in the cardinality attribute of the Attribute class in the frame
model metadata (Fig.
7.17
).
Search WWH ::
Custom Search