Companion Agents: A Table-Information Mining Paradigm for Text-to-SQL

TL;DR

This paper introduces Companion Agents, a database-centric approach that proactively mines and consolidates relational database information to improve Text-to-SQL accuracy in real-world, annotation-scarce scenarios.

Contribution

It proposes a novel paradigm with agents that mine hidden database relations and cues, enhancing Text-to-SQL performance without relying on external knowledge or complete annotations.

Findings

Achieved +4.49 to +14.13 accuracy improvements on multiple datasets.

Significant gains on challenging subsets, up to +16.71 accuracy.

Demonstrated practical applicability for industrial Text-to-SQL deployment.

Abstract

Large-scale Text-to-SQL benchmarks such as BIRD typically assume complete and accurate database annotations as well as readily available external knowledge, which fails to reflect common industrial settings where annotations are missing, incomplete, or erroneous. This mismatch substantially limits the real-world applicability of state-of-the-art (SOTA) Text-to-SQL systems. To bridge this gap, we explore a database-centric approach that leverages intrinsic, fine-grained information residing in relational databases to construct missing evidence and improve Text-to-SQL accuracy under annotation-scarce conditions. Our key hypothesis is that when a query requires multi-step reasoning over extensive table information, existing methods often struggle to reliably identify and utilize the truly relevant knowledge. We therefore propose to "cache" query-relevant knowledge on the database side in advance, so that it can be selectively activated at inference time. Based on this idea, we introduce Companion Agents (CA), a new Text-to-SQL paradigm that incorporates a group of agents accompanying database schemas to proactively mine and consolidate hidden inter-table relations, value-domain distributions, statistical regularities, and latent semantic cues before query generation. Experiments on BIRD under the fully missing evidence setting show that CA recovers +4.49 / +4.37 / +14.13 execution accuracy points on RSL-SQL / CHESS / DAIL-SQL, respectively, with larger gains on the Challenging subset +9.65 / +7.58 / +16.71. These improvements stem from CA's automatic database-side mining and evidence construction, suggesting a practical path toward industrial-grade Text-to-SQL deployment without reliance on human-curated evidence.