KaggleDBQA

该数据集虽然规模较小，但结构复杂，专为真实环境下文本到SQL语义解析器的评估而设计。它包含了8个真实的网络数据库和275个不受限制的问题。在此研究中，KaggleDBQA数据集被用作纯粹的测试集，而Spider的训练集则作为示例库。在规模上，该数据集被归类为小型，其任务是进行文本到SQL的评估。

This dataset, though small in scale, features a complex structure and is specifically designed for evaluating text-to-SQL semantic parsers in real-world scenarios. It contains 8 real-world web databases and 275 unconstrained questions. In this study, the KaggleDBQA dataset is used as a pure test set, while the training split of Spider serves as the example reference corpus. Classified as a small-scale dataset, its core task is text-to-SQL evaluation.