OpenSQL: Data-Efficient Text-to-SQL for Open-Source LLMs via Synthesized Intermediate Supervision

Summary: OpenSQL targets data-efficient Text-to-SQL with open-source LLMs by synthesizing intermediate supervision from scarce (question, SQL) pairs. Global-local schema linking + reasoning-enhanced, clause/semantic-level candidate selection + task-aware augmentation yield 70% Bird-dev with only 14K samples (32B), beating OmniSQL (2.5M). (summarized by gpt-5.4-mini on May 27 2026)

Paper ID: hc4715a7f65bea9a2
Venue: VLDB
Year: 2026
Pagerank: 4.9793485e-05
Overall Rank: 10,738 | 27.81%
DOI: 10.14778/3801059.3801074

Incoming Non-self Citations Over Time

No non-self incoming citations found for this paper in this database.

Authors

1. Ruilin Hu (Tsinghua University)
2. Yuyu Luo (Hong Kong University of Science and Technology)
3. Guoliang Li (Tsinghua University)
4. Shuangqiao Wu (Tencent)
5. Yun Luo (Tencent)

BibTeX Citation

@article{hu_vldb26,
        title = {{OpenSQL: Data-Efficient Text-to-SQL for Open-Source LLMs via Synthesized Intermediate Supervision}},
        author = {Hu, Ruilin and Luo, Yuyu and Li, Guoliang and Wu, Shuangqiao and Luo, Yun},
        journal = {PVLDB},
        series = {{VLDB} '26},
        volume = {19},
        number = {7},
        pages = {1628--1642},
        doi = {10.14778/3801059.3801074},
        url = {https://doi.org/10.14778/3801059.3801074},
        year = {2026}
}

Incoming Citations (Sorted by Pagerank)

Showing 0 of 0 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank

Outgoing Citations (Sorted by Pagerank)

Showing 16 of 16 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
158	Deep Learning for Entity Matching: A Design Space Exploration	2018	SIGMOD	0.00028046388
174	Text-to-SQL Empowered by Large Language Models: A Benchmark Evaluation	2024	VLDB	0.00026790979
377	TURL: Table Understanding through Representation Learning	2021	VLDB	0.00019570264
533	CodeS: Towards Building Open-source Language Models for Text-to-SQL	2024	SIGMOD	0.00016826571
1,356	CatSQL: Towards Real World Natural Language to SQL Applications	2023	VLDB	0.00010925509
1,391	Creating Embeddings of Heterogeneous Relational Datasets for Data Integration Tasks	2020	SIGMOD	0.00010816237
1,777	OmniSQL: Synthesizing High-quality Text-to-SQL Data at Scale	2025	VLDB	9.664552e-05
1,866	ScienceBenchmark: A Complex Real-World Benchmark for Evaluating Natural Language to SQL Systems	2024	VLDB	9.481643e-05
2,025	OpenSearch-SQL: Enhancing Text-to-SQL with Dynamic Few-shot and Consistency Alignment	2025	SIGMOD	9.163622e-05
2,348	The Dawn of Natural Language to SQL: Are We Fully Ready?	2024	VLDB	8.6009821e-05
2,467	Few-shot Text-to-SQL Translation using Structure and Content Prompt Learning	2023	SIGMOD	8.4211526e-05
3,489	Combining Small Language Models and Large Language Models for Zero-Shot NL2SQL	2024	VLDB	7.2627807e-05
4,142	AOP: Automated and Interactive LLM Pipeline Orchestration for Answering Complex Queries	2025	CIDR	6.7814795e-05
6,379	Unify: A System For Unstructured Data Analytics	2025	VLDB	5.8083127e-05
7,952	The Power of Constraints in Natural Language to SQL Translation	2025	VLDB	5.4198293e-05
10,433	DeepEye-SQL: A Software-Engineering-Inspired Text-to-SQL Framework	2026	SIGMOD	4.9793485e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	11,160	RTS+: Reliable Text to SQL	2025	SIGMOD
2	10,832	ReSequel: Robust LLM-assisted Query Rewriting and Optimization using Templatization and Sampling	2026	VLDB
3	10,433	DeepEye-SQL: A Software-Engineering-Inspired Text-to-SQL Framework	2026	SIGMOD
4	3,489	Combining Small Language Models and Large Language Models for Zero-Shot NL2SQL	2024	VLDB
5	7,012	Reliable Text-to-SQL with Adaptive Abstention	2025	SIGMOD
6	10,695	NL2SQLBench: A Modular Benchmarking Framework for LLM-Enabled NL2SQL Solutions	2026	VLDB
7	533	CodeS: Towards Building Open-source Language Models for Text-to-SQL	2024	SIGMOD
8	174	Text-to-SQL Empowered by Large Language Models: A Benchmark Evaluation	2024	VLDB
9	2,025	OpenSearch-SQL: Enhancing Text-to-SQL with Dynamic Few-shot and Consistency Alignment	2025	SIGMOD
10	1,777	OmniSQL: Synthesizing High-quality Text-to-SQL Data at Scale	2025	VLDB