The datasets is that the table corresponding to question is given. But in real industrial application, we have 100+ tables for 1 new question. Thank you! 真正落地的时候，给一个问题，会有100多张表候选，要先定位到表，才能生成SQL，这个有什么方法定位到表？
The datasets is that the table corresponding to question is given.
But in real industrial application, we have 100+ tables for 1 new question.
Thank you!
真正落地的时候，给一个问题，会有100多张表候选，要先定位到表，才能生成SQL，这个有什么方法定位到表？