AV-SQL: Decomposing Complex Text-to-SQL Queries with Agentic Views

2026-04-08

AI Summary

This work addresses the challenge of accurately translating complex natural language queries into SQL over large databases, a task hindered by limited context length and difficulties in multi-table reasoning. The authors propose a multi-agent collaborative framework that decomposes text-to-SQL conversion into three stages: query rewriting, proxy view generation, and SQL synthesis. A key innovation is the introduction of a “proxy view” mechanism, wherein agents dynamically generate common table expressions (CTEs) to compress the database schema, encapsulate intermediate logical steps, and enable stepwise reasoning—thereby mitigating context constraints and schema-linking errors. The approach achieves a state-of-the-art execution accuracy of 70.38% on Spider 2.0 and demonstrates leading performance across multiple benchmarks, including Spider (85.59%), BIRD (72.16%), and KaggleDBQA (63.78%).

Abstract

Text-to-SQL is the task of translating natural language queries into executable SQL for a given database, enabling non-expert users to access structured data without writing SQL manually. Despite rapid advances driven by large language models (LLMs), existing approaches still struggle with complex queries in real-world settings, where database schemas are large and questions require multi-step reasoning over many interrelated tables. In such cases, providing the full schema often exceeds the context window, while one-shot generation frequently produces non-executable SQL due to syntax errors and incorrect schema linking. To address these challenges, we introduce AV-SQL, a framework that decomposes complex Text-to-SQL into a pipeline of specialized LLM agents. Central to AV-SQL is the concept of agentic views: agent-generated Common Table Expressions (CTEs) that encapsulate intermediate query logic and filter relevant schema elements from large schemas. AV-SQL operates in three stages: (1) a rewriter agent compresses and clarifies the input query; (2) a view generator agent processes schema chunks to produce agentic views; and (3) a planner, generator, and revisor agent collaboratively compose these views into the final SQL query. Extensive experiments show that AV-SQL achieves 70.38% execution accuracy on the challenging Spider 2.0 benchmark, outperforming state-of-the-art baselines, while remaining competitive on standard datasets with 85.59% on Spider, 72.16% on BIRD and 63.78% on KaggleDBQA. Our source code is available at https://github.com/pminhtam/AV-SQL.

Problem

• Text-to-SQL
• complex queries
• large database schemas
• multi-step reasoning
• schema linking

Innovation

• Agentic Views
• Text-to-SQL
• Common Table Expressions
• Multi-agent Framework
• Schema Decomposition