The functional data model and the data languages DAPLEX
ACM Transactions on Database Systems (TODS)
Main Memory Orientated Optimization of OO Queries Using Typed Datalog with Foreign Predicates
IEEE Transactions on Knowledge and Data Engineering
Efficient Access to FDM Objects Stored in a Relational Database
BNCOD 12 Proceedings of the 12th British National Conference on Databases: Directions in Databases
Query processing over object views of relational data
The VLDB Journal — The International Journal on Very Large Data Bases
An efficient SQL-based RDF querying scheme
VLDB '05 Proceedings of the 31st international conference on Very large data bases
Rewriting Queries Using Views for RDF-Based Relational Integration
ICTAI '05 Proceedings of the 17th IEEE International Conference on Tools with Artificial Intelligence
SWARD: Semantic Web Abridged Relational Databases
DEXA '07 Proceedings of the 18th International Conference on Database and Expert Systems Applications
ISWC'06 Proceedings of the 5th international conference on The Semantic Web
| Hi-index | 0.00 |
New knowledge can be gained by combining data from various existing autonomous data collections that are held in heterogeneous formats. Relational databases and RDF data are currently two of the most widely used formats for structured data collections supporting user queries. While data from heterogeneous sources could be restructured and imported into a central system for querying this carries a substantial overhead, particularly if the data sources are updated frequently. Therefore, providing support for queries that require access both to existing relational databases and to RDF data collections, without importing large quantities of data, is the problem addressed in this work. To address this problem, we have implemented a federated system in which queries are expressed in a high-level query language against a federated schema that is described using a semantic data model. During processing, a query may be split into several sub-queries that require access to underlying RDFS and relational data resources, and each sub-query is translated to either SPARQL or SQL for execution, thus taking advantage of existing search capabilities and indexes used by those resources, and avoiding the need to import large quantities of data into a central system in order to answer queries. Query splitting and result integration is done in a uniform way that is independent of what kinds of data resources are being accessed. This uniformity simplifies the system's implementation and facilitates modularity in its design.