Search papers, labs, and topics across Lattice.
This paper introduces MDB-Link, a hierarchical schema-linking framework designed to enhance Text-to-SQL performance in multi-database environments by efficiently locating target databases and constructing relevant SQL schemas. The framework employs a global index for retrieving pertinent columns, aggregates evidence for database shortlisting, and utilizes a budget-aware large language model for reranking and selection tasks. MDB-Link significantly improves exact match rates across multiple benchmarks, achieving up to 51.41 on MMQA and outperforming existing methods in both accuracy and speed.
MDB-Link boosts Text-to-SQL accuracy by over 30 points while streamlining the schema selection process in multi-database scenarios.
Traditional Text-to-SQL research and benchmarks assume a known target database, overlooking settings in which a query must be routed within a large, heterogeneous database collection. We therefore study schema linking in a multi-database setting, where the system must first locate the target database and then construct a compact, SQL-relevant schema for generation. We propose MDB-Link, a hierarchical schema-linking framework that retrieves question-relevant columns from a global index, aggregates retrieval evidence to shortlist databases, and uses a budget-aware large language model (LLM) for database reranking, table selection, and column grounding. With Qwen2.5-14B, MDB-Link outperforms LinkAlign on MMQA, Spider2-Snow, and BIRD-dev in database localization and column selection while producing schema subsets close in size to the gold schemas. Exact match improves from 16.88 to 51.41 on MMQA, 2.50 to 9.17 on Spider2-Snow, and 12.52 to 38.01 on BIRD-dev. MDB-Link also runs faster than LinkAlign and AutoLink, demonstrating the effectiveness of hierarchical schema reduction for downstream SQL generation.
