Iteration Without Elaboration: A Simple ReAct Architecture Suffices for Text-to-SQL Generation

arXiv:2608.22651v1 Announce Type: new Abstract: Modern text-to-SQL systems have become increasingly elaborate, relying on schema-linking modules, retrieval-augmented prompting, candidate generation, and multi-stage refinement pipelines. While effective, these additions introduce substantial latency and engineering overhead. To this end, we present extbf{ReAct-SQL}, a simple yet effective zero-shot ReAct-style framework built solely on iterative reasoning and a constrained action space defined by a typed Domain-Specific Language (DSL) of 15 relational operations, rather than free-form SQL generation. The model incrementally issues DSL calls, observes compiled-SQL execution feedback, and revises its reasoning through interaction. On corrected BIRD mini-dev and EHR-SQL, ReAct-SQL achieves extbf{84.5%} and extbf{73.9%} accuracy, respectively, matching substantially more elaborate baselines while running up to 8imes faster. Incremental ablations further show that iteration primarily improves grounding, while the DSL improves compositional reliability.

Source: arXiv cs.CL | 2026-08-25