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

TL;DR

This paper introduces KaSLA, a knapsack optimization-based schema linking method that enhances SQL generation from user queries by accurately identifying relevant schema elements and reducing redundancy.

Contribution

KaSLA is a novel plug-in schema linking approach that employs hierarchical linking and knapsack optimization to improve schema linking accuracy for Text2SQL tasks.

Findings

KaSLA outperforms existing schema linking methods on Spider and BIRD benchmarks.

KaSLA significantly improves SQL generation performance of state-of-the-art Text2SQL models.

The code for KaSLA is publicly available at the provided GitHub link.

Abstract

Generating SQLs from user queries is a long-standing 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 \textbf{restricted missing indicator}. Accordingly, we introduce \textbf{\underline{K}n\underline{a}psack optimization-based \underline{S}chema \underline{L}inking \underline{A}pproach (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-the-art (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.