Virtual Ontology Layer for SQL Semantic Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Relational databases struggle to capture and extract contextual and semantic knowledge from stored data due to their limitations in representing inherent meaning, requiring advanced query languages like SPARQL that are beyond the skills of conventional database users.
Innovation Solution
A system and method for implementing a virtual ontology over a relational database, which defines a virtual ontology by declaring a hierarchy of virtual entities, mapping them to physical tables, and storing metadata in a virtual ontology database. This allows querying the ontology metadata database to rewrite relational queries and extract ontological knowledge within a relational framework.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If relational databases are used to store data, then data organization and retrieval are simplified, but the ability to capture and extract semantic knowledge is limited
Solution Approach 1:
The patent segments the database system into two distinct layers: a relational database layer for storing data in traditional tables, and a virtual ontology layer for representing semantic knowledge. This segmentation allows each layer to serve its specific purpose - the relational layer maintains simple data organization while the ontology layer captures semantic relationships, resolving the contradiction between ease of operation and semantic knowledge preservation.
Solution Approach 2:
The patent introduces a virtual ontology as an intermediary between the relational database and the user. This virtual ontology acts as a mediator that maps relational data to semantic concepts, enabling users to query semantic knowledge without directly interacting with the underlying relational schema. The intermediary layer translates between relational and semantic query languages, preserving both the simplicity of relational access and the richness of semantic knowledge.
2Loss of information
If advanced query languages like SPARQL are used to extract semantic knowledge, then semantic capabilities are improved, but the complexity of operation increases
Solution Approach 1:
The patent creates a virtual ontology that is a conceptual copy or representation of the relational data structured in semantic terms. This virtual ontology can be queried using familiar SQL-like syntax rather than complex SPARQL, providing an accessible interface that copies the simplicity of relational queries while enabling semantic knowledge extraction through the virtual ontology layer.
Solution Approach 2:
The virtual ontology serves as an intermediary that translates between simple SQL queries and complex semantic queries. Users can write queries in familiar SQL syntax against the virtual ontology, and the system automatically translates these into the appropriate semantic queries to extract knowledge from the relational database, thereby maintaining ease of operation while achieving semantic capabilities.
3Loss of information
If a virtual ontology is implemented over a relational database, then semantic knowledge extraction is enabled, but the system complexity increases
Solution Approach 1:
The virtual ontology is implemented as a universal layer that can represent multiple different ontologies and semantic models over the same relational database. This multi-functionality allows the system to provide semantic knowledge extraction for various domains without requiring separate complex systems for each ontology, thereby managing complexity through a single versatile framework.
Solution Approach 2:
The patent performs preliminary actions by pre-defining the virtual ontology structure, mappings between relational tables and ontology concepts, and transformation rules before actual queries are executed. This preliminary setup of the ontology framework and its relationships with the relational database reduces the complexity of individual queries, as the heavy lifting of semantic interpretation is prepared in advance.
Data Source
AI summary
A method for implementing a virtual ontology over a relational database, comprising: defining a virtual ontology from multiple candidates for a relational database by: declaring a hierarchy of virtual entities; for each virtual entity, declaring a mapping associating the virtual entity with at least a subset of a physical table in the database; storing the virtual entity in association with the mapping in a metadata database; and defining multiple schemas, each implemented as a different relational view of the virtual ontology and capturing at least one ontology capability, wherein querying the metadata database with a first relational query targeting one of the schemas enables writing a second relational query, wherein executing the second relational query at the relational database extracts knowledge that is opaque to the relational schema and expresses the ontology capability captured by the targeted ontology schema, thereby extracting the knowledge within a relational framework.