Text2sql-Flow: a Robust Sql-Aware Data Augmentation Framework for Text-To-Sql

The data-centric paradigm has emerged as a pivotal direction in artificial intelligence (AI), relying on high-quality training data. This shift is especially critical in the Text-to-SQL task, where the scarcity, limited diversity, and structural simplicity of existing datasets constrain model performance. To address these challenges, we propose TEXT2SQL-FLOW, a SQLaware data augmentation framework that systematically generates large-scale, semantically valid, and structurally diverse Textto-SQL pairs from limited seed data. Our framework operates along six augmentation dimensions and integrates an end-to-end pipeline featuring auxiliary database selection, SQL executability verification, natural language (NL) question generation, NL-SQL correspondence verification, and chain-of-thought (CoT) reasoning trace generation. Leveraging this framework, we construct SQLFLOW, a high-quality dataset comprising 75,386 annotated examples. We demonstrate the utility of SQLFLOW in both finetuning and prompt-based settings: (1) For open-source large language models (LLMs), fine-tuning with SQLFLOW enhances the problem-solving capabilities. Under the same data budget, models trained on SQLFLOW achieve competitive performance gains across multiple benchmarks. (2) For closed-source LLMs, we propose a masked alignment retrieval method that leverages SQLFLOW as both a knowledge base and the training data for the retrieval model. This approach enables structure-aware example matching by modeling fine-grained alignments between NL questions and SQL queries. Experimental results show that our retrieval strategy outperforms existing example retrieval methods, highlighting the dual importance of SQLFLOW's highquality data and our novel retrieval technique. Our work establishes a scalable, data-centric foundation for advancing Textto-SQL systems and underscores the indispensable role of structured, high-fidelity data in modern AI development. Our code is available at https://github.com/TechNomad-ds/Text2SQL-Flow.