Salvato in:
Dettagli Bibliografici
Autori principali: Perlitz, Yotam, Venezian, Elad, Royer, Corentin, Fusco, Francesco, Giovannini, Andrea
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:https://arxiv.org/abs/2605.07796
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910201540509696
author Perlitz, Yotam
Venezian, Elad
Royer, Corentin
Fusco, Francesco
Giovannini, Andrea
author_facet Perlitz, Yotam
Venezian, Elad
Royer, Corentin
Fusco, Francesco
Giovannini, Andrea
contents SQL dialects vary in syntax, types, and functions across database engines. Text-to-SQL benchmarks, however, predominantly support only SQLite. This creates a critical evaluation gap: cross-dialect evaluation reveals weak per-query agreement (Cohen's ), showing that SQLite performance is an unreliable proxy for other dialects. Yet such evaluation remains prohibitively difficult: existing approaches either require expensive manual query transpilation or rely on tools that often fail on complex SQL. To close this gap, we introduce PolySQL, a novel dual-execution method that eliminates the need for query transpilation by comparing normalized execution results. Notably, our approach achieves higher evaluation fidelity than query transpilation with 100% query coverage. PolySQL comprises three datasets, enabling the first large-scale cross-dialect study. Our study reveals a 10.1% average accuracy drop from SQLite to other dialects and identifies a significant dialect difficulty hierarchy. We find this degradation stems from logical rather than syntactic errors (61% vs. 8%). We release our framework code and leaderboard to enable rigorous dialect-robust evaluation.
format Preprint
id arxiv_https___arxiv_org_abs_2605_07796
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle PolySQL: Scaling Text-to-SQL Evaluation Across SQL Dialects via Automated Backend Isomorphism
Perlitz, Yotam
Venezian, Elad
Royer, Corentin
Fusco, Francesco
Giovannini, Andrea
Computation and Language
SQL dialects vary in syntax, types, and functions across database engines. Text-to-SQL benchmarks, however, predominantly support only SQLite. This creates a critical evaluation gap: cross-dialect evaluation reveals weak per-query agreement (Cohen's ), showing that SQLite performance is an unreliable proxy for other dialects. Yet such evaluation remains prohibitively difficult: existing approaches either require expensive manual query transpilation or rely on tools that often fail on complex SQL. To close this gap, we introduce PolySQL, a novel dual-execution method that eliminates the need for query transpilation by comparing normalized execution results. Notably, our approach achieves higher evaluation fidelity than query transpilation with 100% query coverage. PolySQL comprises three datasets, enabling the first large-scale cross-dialect study. Our study reveals a 10.1% average accuracy drop from SQLite to other dialects and identifies a significant dialect difficulty hierarchy. We find this degradation stems from logical rather than syntactic errors (61% vs. 8%). We release our framework code and leaderboard to enable rigorous dialect-robust evaluation.
title PolySQL: Scaling Text-to-SQL Evaluation Across SQL Dialects via Automated Backend Isomorphism
topic Computation and Language
url https://arxiv.org/abs/2605.07796