Using Knowledge Graphs to Generate SQL Queries from Textual Specifications

Using Knowledge Graphs to Generate SQL Queries from Textual Specifications

연구 분야: Databases

학회: International Conference on Conceptual Modeling

In this paper, we present a tool for querying relational DBs that uses a KG as an approach to generate SQL queries from NL specifications. In this approach, we argue that a KG representation of a relational DB schema can become an auxiliary tool in the translation process. Furthermore, we propose to automate the process of generating such a KG. Our approach to provide an NL interface for relational DBs comprises two major tasks: (1) generation of a KG from a relational DB schema and (2) translation of NL queries to SQL based on the semantics provided by the respective KG. We study the effectiveness of our approach using a benchmark dataset containing 82 NL query examples from the Spider dataset, considering the domain of Formula 1. Our approach is able to correctly translate these queries, which is verified against the expected results provided by our benchmark.