Automatic Database View Generation via Path Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Developing views in large and complex enterprise data models is time-consuming and expensive due to the extensive documentation required, making it cumbersome to create interfaces or reports, and existing solutions lack efficiency in leveraging existing view knowledge for new view generation.
Innovation Solution
A system for automatic view generation that identifies and combines existing database object identifiers to create new views by calculating scores for paths and selecting the most effective ones, either complete or partial, to generate requested views, thereby reducing development time and improving view quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If extensive documentation is used to document large enterprise data models, then completeness of documentation is improved, but ease of operation deteriorates due to cumbersome referencing
Solution Approach 1:
The system creates simplified copies of data model information in the form of graphical icons representing database objects. These icons are displayed in a visual canvas, allowing users to interact with simplified representations rather than navigating extensive textual documentation, thus maintaining information accessibility while dramatically improving ease of operation
Solution Approach 2:
The system transitions from traditional linear textual documentation to a two-dimensional graphical interface where database objects are represented as icons in a visual canvas. This dimensional change allows users to perceive and navigate data model relationships spatially rather than sequentially, improving both accessibility and ease of operation
2Manufacturing precision
If manual view development is performed in large data models, then view quality can be controlled, but productivity deteriorates due to time-consuming development
Solution Approach 1:
The system performs preliminary actions by automatically generating candidate views using AI/ML models trained on existing data model knowledge and usage patterns. These pre-generated candidates are then presented to users for selection and refinement, combining automated efficiency with human quality control to simultaneously improve productivity and maintain view quality
Solution Approach 2:
The system enables self-service view generation by allowing users to simply specify desired data objects and relationships, then the AI/ML engine automatically generates appropriate views without requiring users to manually construct complex queries or navigate documentation. This self-service approach dramatically accelerates productivity while the AI ensures quality through learned best practices
3Productivity
If existing view knowledge is leveraged for new view generation, then productivity is improved, but device complexity increases due to AI/ML integration
Solution Approach 1:
The system introduces an AI/ML-based intermediary layer that sits between the user's view requirements and the database system. This intermediary automatically translates high-level user specifications into detailed view definitions by leveraging learned patterns from existing views, thereby improving productivity while encapsulating complexity within the AI layer rather than exposing it to users
4Manufacturing precision
If comprehensive path scoring is performed to select optimal paths, then view quality is improved, but loss of time increases due to calculation overhead
Solution Approach 1:
The system changes parameters by using AI/ML models to pre-calculate and store scoring metrics for different path options during system initialization or caching phases. When generating views, the system retrieves and compares these pre-computed scores rather than performing exhaustive real-time calculations, thereby maintaining high path selection quality while significantly reducing calculation time
Data Source
AI summary
A process for mining existing views to generate new views is provided herein. A request to generate a proposed view combining a plurality of tables in a database may be received. One or more existing views in the database applicable to the request may be identified. The applicable existing views may include the plurality of tables in the request. One or more paths combining the plurality of tables may be identified in the respective one or more existing views. The identified one or more paths may be provided. One or more scores for the one or more paths may be generated and provided along with the one or more paths. A path may be automatically selected based on the scores or a path may be selected by a user. A view may be generated based on the selected path.