SEDE
收藏资源简介:
SEDE数据集是由哥伦比亚大学创建的,包含12,023对自然语言和SQL查询的配对,这些数据是从Stack Exchange网站的真实用户交互中收集的。数据集内容丰富,涵盖了多种实际挑战,如参数使用、日期操作和复杂的子查询等。创建过程中,研究人员通过规则过滤和人工验证确保数据质量。SEDE数据集主要用于评估和改进Text-to-SQL模型在真实世界环境中的性能,特别是在处理自然语言查询和复杂SQL结构时的能力。
The SEDE dataset was developed by Columbia University, consisting of 12,023 paired natural language and SQL queries collected from real user interactions across the Stack Exchange website. The dataset encompasses a diverse set of practical challenges, such as parameter utilization, date manipulation, complex subqueries and more. During its curation, researchers employed rule-based filtering and manual verification to ensure data quality. The SEDE dataset is primarily used to evaluate and improve the performance of Text-to-SQL models in real-world scenarios, particularly their ability to handle natural language queries and complex SQL structures.




