Global Dictionary Compression for Unified Knowledge Graph Queries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Maintaining multiple data systems for traditional databases and RDF data models incurs high maintenance costs and requires separate queries and additional processing, making it inefficient to handle data in different formats.
Innovation Solution
A triple integration and querying system (TQS) integrates RDF-formatted data into a relational database using a table designed for triples, enabling a single query to access both traditional and RDF data, and employs dictionaries to reduce memory overhead and facilitate efficient querying.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple data systems are used to maintain traditional database data and RDF data separately, then data format integrity is preserved, but maintenance costs increase and system complexity increases
Solution Approach 1:
The patent combines traditional database data and RDF data into a single unified data system. The columnar storage structure accommodates both SQL-formatted data and RDF triples (subject, predicate, object) in the same table format, eliminating the need for separate storage systems while maintaining data integrity through a global dictionary that handles both data types consistently.
Solution Approach 2:
The unified table structure serves multiple functions: it stores both traditional relational data and RDF knowledge graph data, supports both SQL and SPARQL query languages, and uses a single global dictionary for both data types. This multi-functional design reduces system complexity while maintaining compatibility with both data models.
2Reliability
If multiple data systems are used to maintain different data formats, then data type specificity is maintained, but maintenance costs increase
Solution Approach 1:
The patent merges multiple maintenance operations into a single system. The global dictionary is built once from all unique values across both SQL and RDF data, and single maintenance operations can update the unified structure. This eliminates the need for parallel maintenance of separate systems while preserving data type specificity through the dictionary's value-to-integer mappings.
3Measurement precision
If separate queries are used for traditional database and RDF data, then query language specialization is maintained, but processing time increases
Solution Approach 1:
The unified table structure enables a single query operation to access both traditional database data and RDF knowledge graph data. The global dictionary provides a common interface that allows SQL queries and SPARQL queries to operate on the same underlying data structure, eliminating the need for separate query processing paths and reducing overall processing time.
4Reliability
If multiple data systems are used to store different data formats, then data model integrity is preserved, but resource consumption increases
Solution Approach 1:
The patent combines multiple data storage systems into a single columnar storage structure that handles both SQL and RDF data. This consolidation reduces the total computational resources required for data management, as a single system can optimize storage and retrieval operations for both data types simultaneously rather than maintaining separate optimized systems.
Data Source
AI summary
Various embodiments for a triple integration and querying system with dictionary compression are described herein. An embodiment operates by identifying a table of a database with four or more columns with triple formatted data including one subject column, one predicate column, and two or more object columns. It is determined that a master dictionary is to be generated for the both the subject column and the predicate column based on an identical datatype being used for both columns. A master dictionary including both the unique values from the subject data dictionary and the predicate data dictionary is generated. Values in the subject column and the predicate column are replaced based on the unique values from the master dictionary.