Combining the Best of Global-as-View and Local-as-View for Data Integration

1
Combining the Best of Global-as-View and
Local-as-View for Data Integration

• Li Xu
• Brigham Young University

Funded by NSF
2
Data Integration

• A Global Schema
• Global-as-View (GaV) vs. Local-as-View (LaV)
• Query Reformulation
• Mapping
• Adding New Sources

3
Global as View (GaV)
Mediation Description Car has Year (x, y) -
s1.Car has Year(x, y) Car has Feature(x, y) -
s1.Car has Feature(x, y) Car has Mileage(x, y) -
s1.Car has Miles(x, y) Car has MakeModel(x, y)
- s1.Car has MakeModel(x, y)
Year
Feature
Make
Model
Phone
Mileage
4
Local as View (GaV)
Year
Feature
Make
Model
Source Description s1.Car has Year (x, y) -
Car has Year(x, y) s1.Car has Miles (x, y) - Car
has Mileage(x, y) S1.Car has Feature(x, y) - Car
has Feature(x, y)
Phone
Mileage
5
Target-based Integration Query System (TIQS)

• Combining the Best of GaV and LaV
• Rule Unfolding
• Scalability
• Schema Matching
• Source-to-Target Mappings
• Automating Mediation Description

6
Schema Matching

• Mapping Elements
• Direct Matches
• Indirect Matches
• Manipulation Operations
• Mapping Algebra

7
Source-to-Target Mapping
Color
Year
Year
Feature
Make
Feature
Make Model
Model
Body Type
Style
Miles
Phone
Mileage
Source
8
Source-to-Target Mapping (Cont.)
Color
Year
Feature
Body Type
Car
Style
Miles
Source Evaluation
9
Source-to-Target Mapping (Cont.)
Color
Year
Feature
Body Type
Car
Style
Miles
Source Evaluation
10
Source-to-Target Mapping (Cont.)
Year
Year
Feature
Make
Mediation Description Car has Year (x, y) -
s1.Car has Year(x, y) Car has Feature(x, y) -
s1.Car has Feature(x, y) Car has Make(x, y) -
s1.Car has Make(x, y) Car has Model(x, y) -
s1.Car has Model(x, y) Car has Mileage(x, y) -
s1.Car has Miles(x, y)
Model
Car
Phone
Mileage
Miles
Source
11
Query Processing

• User Queries Logic Rules
• Conjunctive Query
• Conjunctive Query with Arithmetic Comparison
• Recursive Query
• Theorem Query Answers
• Sound
• Maximal

12
Conclusion

• A Flexible and Scalable Data Integration Approach
• A Practical Approach
• A Correlation of Schema Matching and Data
  Integration