Database Query Translation via Restricted Schema Views
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users of database management systems lack a simple and flexible method to access information without needing to know the database structure or specific query languages, as existing solutions like APIs require extensive learning and are rigid, making it difficult to formulate queries across multiple database tables.
Innovation Solution
A method that converts user requests in a simple language into a second language to query a database, using a restricted representation of the database to simplify the querying process, allowing for dynamic adaptation to the database structure and optimizing the query by focusing on the objects and properties of interest, without requiring in-depth knowledge of the database structure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If users query the database using its own language (SQL) and know the database structure, then the query can be formulated accurately, but the complexity of operation increases and requires in-depth knowledge
Solution Approach 1:
The patent introduces an intermediary layer (query processing system) that translates natural language queries into database-specific queries. This mediator handles the complexity of database structure and language, allowing users to query without knowing SQL or the underlying database schema, thus resolving the contradiction between query accuracy and operational simplicity
Solution Approach 2:
The system creates a simplified copy or view of the database structure that maps natural language concepts to database objects. This abstraction layer allows users to interact with a simplified representation rather than the complex actual database structure, maintaining query accuracy while reducing operational complexity
2Ease of operation
If libraries and APIs are implemented to simplify database access, then ease of operation improves, but the solution becomes rigid and requires extensive learning
Solution Approach 1:
The patent implements a dynamic query processing system that adapts to different database structures and user needs rather than relying on fixed APIs. The system dynamically generates appropriate queries based on the actual database schema and the user's natural language input, providing simplicity without the rigidity and learning curve of predefined APIs
Solution Approach 2:
The system changes the parameters of interaction by accepting natural language input with varying levels of specificity and automatically adjusting the query formulation based on the database structure. This flexibility eliminates the need for users to learn complex API parameters while maintaining ease of operation
3Adaptability or versatility
If the complete database structure is exposed to users, then all information can be accessed, but the complexity of querying increases
Solution Approach 1:
The patent extracts only the necessary portions of the database structure relevant to the user's query rather than exposing the complete structure. The query processing system identifies and extracts only the tables, columns, and relationships needed to answer the specific question, reducing complexity while maintaining full information accessibility when needed
4Loss of information
If multiple tables are queried simultaneously for comprehensive information, then the completeness of results improves, but the difficulty of formulating the query increases
Solution Approach 1:
The query processing system uses feedback from the database structure analysis to automatically determine which tables need to be joined and how. The system analyzes the user's natural language query, feedback from the database schema, and automatically constructs the appropriate multi-table query, eliminating the difficulty of manually formulating complex joins while ensuring information completeness
Data Source
Figure 1~4
Figure 2
Figure 3
AI summary
The invention relates to a method for querying a database by a computer means, comprising the following steps: receiving a first request in a first language, aimed at obtaining information contained in the database, said request being aimed at least at one object having at least one property represented in the database; accessing a representation of the database restricted to said at least one object; generating, from said representation and said at least one property of the first request, at least one second request in a second language aimed at obtaining the information requested in the first request; and querying the database by means of said at least one second request.