Semantic-to-Graph Model Inference Without ETL Data Copying
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for constructing graph models are resource-intensive and require multiple data copies, leading to challenges in data management, accuracy, privacy, and security.
Innovation Solution
A method to infer a graph model directly from a semantic model, utilizing a key identifying component to identify primary keys based on relationship types, creating vertex and edge tables, and generating a graph model over existing data without copying, thereby reducing resource consumption and enabling quick deployment of graph queries and algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If ETL processes are used to construct a graph database by copying source data, then a complete graph database can be created for data analysis queries, but computing resources and storage space are significantly consumed
Solution Approach 1:
The patent extracts only the essential schema information (tables, columns, relationships, primary keys, foreign keys) from the source database metadata rather than copying actual data. This creates a lightweight graph model that preserves structural relationships while eliminating the need to store and process duplicate data, significantly reducing computing resource consumption.
Solution Approach 2:
The patent creates a virtual copy of the database schema structure through graph model representation, where vertices represent tables and edges represent relationships. This structural copy enables graph queries without requiring physical data duplication, maintaining graph database functionality while minimizing resource usage.
2Reliability
If ETL processes are used to construct a graph database by copying source data, then a graph database can be created for data analysis, but human resources for designing the graph database are significantly consumed
Solution Approach 1:
The system automatically generates the graph model by extracting metadata from the source database and inferring relationships based on primary keys and foreign keys. This self-service approach eliminates the need for manual graph database design by data engineers, automatically creating vertices, edges, and relationship structures without human intervention in the design process.
Solution Approach 2:
The patent performs preliminary metadata extraction and relationship inference during the model creation phase, establishing the complete graph structure before any queries are executed. This preliminary action includes identifying all tables as vertices, all relationships as edges, and establishing the directional connections based on foreign key references, thereby eliminating subsequent design iterations.
3Adaptability or versatility
If ETL processes create multiple copies of data, then data can be accessed from multiple sources, but data accuracy, privacy, and security management becomes complex
Solution Approach 1:
The patent introduces a metadata-based graph model as an intermediary layer between the source database and query interfaces. This intermediary represents data relationships without creating additional data copies, allowing flexible data access through graph queries while maintaining a single source of truth. The graph model acts as a virtual schema that enables diverse query patterns without duplicating underlying data.
Solution Approach 2:
The patent creates a virtual copy of the database schema through graph representation, where the graph structure mirrors the source database relationships without duplicating actual data values. This schema-level copying enables multiple access paths and query types while maintaining data integrity and security through the single source database.
Data Source
AI summary
Apparatuses, methods, and computer-readable media infer a graph model from a semantic model that describes source tables and relationships among the source tables. The apparatus identifies at least one primary key for each relationship of a plurality of relationships defined in the semantic model based on whether the relationship is a m:1 relationship or a 1:1 relationship, each relationship defining at least one from-table and one to-table. The apparatus creates a vertex table in the graph model for each source table having a primary key. The apparatus creates an edge table in the graph model for each from-table defined in a m:1 relationship of the semantic model and creates an edge table in the graph model for one of the from-table or the to-table in a 1:1 relationship of the semantic model. Each edge of the edge table is labeled with a source or destination vertex table and a foreign-key.