Knapsack Optimization-Based Schema Linking for LLM-Based Text-to-SQL Generation

Generating SQLs from user queries is a longstanding challenge, where the accuracy of initial schema linking significantly impacts subsequent SQL generation performance. However, current schema linking models still struggle with missing relevant schema elements or an excess of redundant ones. A crucial reason for this is that commonly used metrics, recall and precision, fail to capture relevant element missing and thus cannot reflect actual schema linking performance. Motivated by this, we propose enhanced schema linking metrics by introducing a restricted missing indicator. Accordingly, we introduce Knapsack optimization-based Schema Linking Approach (KaSLA), a plug-in schema linking method designed to prevent the missing of relevant schema elements while minimizing the inclusion of redundant ones. KaSLA employs a hierarchical linking strategy that first identifies the optimal table linking and subsequently links columns within the selected table to reduce linking candidate space. In each linking process, it utilizes a knapsack optimization approach to link potentially relevant elements while accounting for a limited tolerance of potentially redundant ones. With this optimization, KaSLA-1.6B achieves superior schema linking results compared to large-scale LLMs, including DeepSeek-V3 with the state-of-theart (SOTA) schema linking method. Extensive experiments on Spider and BIRD benchmarks verify that KaSLA can significantly improve the SQL generation performance of SOTA Text2SQL models by substituting their schema linking processes. The code is available at https://github.com/DEEP-PolyU/KaSLA.

Index Terms-text-to-SQL, database, large language models, natural language understanding a significant 14.91% performance gap between state-of-theart schema linking methods and ground-truth linking results. Consequently, enhancing schema linking accuracy represents a critical research frontier with significant potential to advance text-to-SQL capabilities.

Recent state-of-the-art text-to-SQL models primarily focus on the final SQL generation stage, often relying on basic or even neglecting schema linking strategies. Existing schema