RASL: Retrieval Augmented Schema Linking for Massive Database Text-to-SQL
It addresses scalability for enterprise database interfaces, enabling practical deployment across diverse settings without specialized tuning, though it appears incremental as it builds on existing retrieval and schema linking approaches.
The paper tackles the challenge of scaling text-to-SQL systems to massive enterprise databases by introducing a retrieval-augmented schema linking method that decomposes schemas into semantic units for targeted retrieval, resulting in high recall and accuracy outperforming baselines without requiring fine-tuning.
Despite advances in large language model (LLM)-based natural language interfaces for databases, scaling to enterprise-level data catalogs remains an under-explored challenge. Prior works addressing this challenge rely on domain-specific fine-tuning - complicating deployment - and fail to leverage important semantic context contained within database metadata. To address these limitations, we introduce a component-based retrieval architecture that decomposes database schemas and metadata into discrete semantic units, each separately indexed for targeted retrieval. Our approach prioritizes effective table identification while leveraging column-level information, ensuring the total number of retrieved tables remains within a manageable context budget. Experiments demonstrate that our method maintains high recall and accuracy, with our system outperforming baselines over massive databases with varying structure and available metadata. Our solution enables practical text-to-SQL systems deployable across diverse enterprise settings without specialized fine-tuning, addressing a critical scalability gap in natural language database interfaces.