UNITE

UNITE是一个综合性的文本到SQL评估基准，由AWS AI Labs创建，旨在全面评估文本到SQL系统的性能。该数据集包含来自超过12个领域的自然语言问题，涵盖了3.9K种SQL查询模式和29K个数据库，总计有97K训练和27K测试示例。UNITE通过统一格式，促进了文本到SQL的通用性和鲁棒性的基准测试。数据集的应用领域广泛，旨在解决文本到SQL系统在实际应用中的通用性和鲁棒性问题。

UNITE is a comprehensive text-to-SQL evaluation benchmark developed by AWS AI Labs, which aims to thoroughly assess the performance of text-to-SQL systems. This dataset comprises natural language questions from over 12 domains, covers 3.9K SQL query patterns and 29K databases, and includes a total of 97K training and 27K test instances. By adopting a unified format, UNITE facilitates benchmarking research on the generality and robustness of text-to-SQL systems. With broad application domains, this dataset is designed to address the issues of generality and robustness of text-to-SQL systems in real-world applications.