arXiv 2508.127692025-08-20cs.CLcs.AI

CRED-SQL：通过聚类检索和执行描述增强真实世界大规模数据库文本转SQL解析

CRED-SQL: Enhancing Real-world Large Scale Database Text-to-SQL Parsing through Cluster Retrieval and Execution Description

Harbin Institute of Technology, Shenzhen, China（哈尔滨工业大学（深圳））；Pengcheng Laboratory, Shenzhen, China（鹏城实验室）；Inspur Cloud Information Technology Co., Ltd, Jinan 250101, China（浪潮云信息技术有限公司）；Guangdong Power Grid Co., Ltd, China（广东电网公司）

Shaoming Duan, Zirui Wang, Chuanyi Liu, Zhibin Zhu, Yuhao Zhang, Peiyi Han, Liang Yan, Zewu Peng

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

针对大规模数据库中文本转SQL任务存在的语义不匹配问题，提出集成聚类检索和执行描述的CRED-SQL框架，通过聚类模式检索和EDL中间表示提升性能，在SpiderUnion和BirdUnion上实现SOTA。

摘要：

大型语言模型（LLMs）的最新进展显著提高了文本转SQL系统的准确性。然而，一个关键挑战仍然存在：自然语言问题（NLQs）与其对应的SQL查询之间存在语义不匹配。这一问题在大规模数据库中更为严重，其中语义相似的属性会阻碍模式链接，且SQL生成过程中的语义漂移最终会降低模型准确性。为应对这些挑战，我们引入CRED-SQL，一个专为大规模数据库设计的框架，集成了聚类检索和执行描述。CRED-SQL首先执行基于聚类的大规模模式检索，以精确定位与给定NLQ相关的表和列，缓解模式不匹配问题；然后引入一种中间自然语言表示——执行描述语言（EDL），以弥合NLQ与SQL之间的差距。这种重构将任务分解为两个阶段：文本转EDL和EDL转SQL，既利用了LLMs强大的通用推理能力，又减少了语义偏差。在两个大规模跨域基准数据集SpiderUnion和BirdUnion上的大量实验表明，CRED-SQL实现了新的最先进（SOTA）性能，验证了其有效性和可扩展性。我们的代码可在https://github.com/smduan/CRED-SQL.git获取。

英文摘要：

Recent advances in large language models (LLMs) have significantly improved the accuracy of Text-to-SQL systems. However, a critical challenge remains: the semantic mismatch between natural language questions (NLQs) and their corresponding SQL queries. This issue is exacerbated in large-scale databases, where semantically similar attributes hinder schema linking and semantic drift during SQL generation, ultimately reducing model accuracy. To address these challenges, we introduce CRED-SQL, a framework designed for large-scale databases that integrates Cluster Retrieval and Execution Description. CRED-SQL first performs cluster-based large-scale schema retrieval to pinpoint the tables and columns most relevant to a given NLQ, alleviating schema mismatch. It then introduces an intermediate natural language representation-Execution Description Language (EDL)-to bridge the gap between NLQs and SQL. This reformulation decomposes the task into two stages: Text-to-EDL and EDL-to-SQL, leveraging LLMs' strong general reasoning capabilities while reducing semantic deviation. Extensive experiments on two large-scale, cross-domain benchmarks-SpiderUnion and BirdUnion-demonstrate that CRED-SQL achieves new state-of-the-art (SOTA) performance, validating its effectiveness and scalability. Our code is available at https://github.com/smduan/CRED-SQL.git

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。