TUR2SQL dataset consists of 10,809 pairs of natural language statements and their corresponding SQL queries. The generation of TUR2SQL is detailed in our UBMK-2023 paper:
@inproceedings{tur2sql,
title = {TUR2SQL: A Cross-Domain Turkish Dataset For Text-to-SQL},
author = {Kanburoğlu, Ali Buğra and Tek, F. Boray},
booktitle = {2023 8th International Conference on Computer Science and Engineering (UBMK)},
year = {2023},
pages = {206-211},
doi = {10.1109/UBMK59864.2023.10286686}
}