Page 178 - 《软件学报》2026年第3期

刘孟怡 等: 基于大语言模型的空间数据库自然语言查询转换方法 1141

Workshop on Spatio-temporal Reasoning and Learning Co-located with the 32nd Int’l Joint Conf. on Artificial Intelligence. 2023.
[20] Wang XY, Liu MY, Xu JQ, Lu H. NALMO: Transforming queries in natural language for moving objects databases. GeoInformatica,
2023, 27(3): 427–460. [doi: 10.1007/S10707-023-00494-5]
[21] Fan YK, He ZY, Ren TH, Guo DJ, Chen L, Zhu RS, Chen GD, Jing YN, Zhang K, Wang XS. GAR: A generate-and-rank approach for
natural language to SQL translation. In: Proc. of the 39th IEEE Int’l Conf. on Data Engineering. Anaheim: IEEE, 2023. 110–122.
[doi: 10.1109/ICDE55515.2023.00016]
[22] Lehner B, Verdin K, Jarvis A. New global hydrography derived from spaceborne elevation data. Eos, Trans., American Geophysical
Union, 2008, 89(10): 93–94. [doi: 10.1029/2008EO100001]
[23] Wang WL, Li JJ, Ku WS, Wang HX. Multilingual spatial domain natural language interface to databases. GeoInformatica, 2024, 28(1):
29–52. [doi: 10.1007/S10707-023-00496-3]
[24] Sen J, Lei C, Quamar A, Özcan F, Efthymiou V, Dalmia A, Stager G, Mittal A, Saha D, Sankaranarayanan K. ATHENA++: Natural
language querying for complex nested SQL queries. Proc. of the VLDB Endowment, 2020, 13(12): 2747–2759. [doi: 10.14778/3407790.
3407858]
[25] Gao DW, Wang HB, Li YL, Sun XY, Qian YC, Ding BL, Zhou JR. Text-to-SQL empowered by large language models: A benchmark
evaluation. Proc. of the VLDB Endowment, 2024, 17(5): 1132–1145. [doi: 10.14778/3641204.3641221]
[26] Sun S, Zhang YC, Yan JH, Gao YZ, Ong D, Chen B, Su J. Battle of the large language models: Dolly vs LLaMA vs Vicuna vs Guanaco
vs Bard vs ChatGPT—A text-to-SQL parsing comparison. In: Findings of the Association for Computational Linguistics: EMNLP 2023.
Singapore: ACL, 2023. 11225–11238. [doi: 10.18653/V1/2023.FINDINGS-EMNLP.750]

附中文参考文献
[1] 刘志丹, 林维鑫, 伍楷舜. 一种面向大规模空间文本数据的空间结构匹配算法. 计算机学报, 2022, 45(6): 1261–1275. [doi: 10.
11897/SP.J.1016.2022.01261]
[12] 李诗晨, 王中卿, 周国栋. 大语言模型驱动的跨领域属性级情感分析. 软件学报, 2025, 36(2): 644–659. http://www.jos.org.cn/1000-
9825/7156.htm [doi: 10.13328/j.cnki.jos.007156]
[16] 潘璇, 徐思涵, 蔡祥睿, 温延龙, 袁晓洁. 基于深度学习的数据库自然语言接口综述. 计算机研究与发展, 2021, 58(9): 1925–1950.
[doi: 10.7544/issn1000-1239.2021.20200209]

附录 A. 可执行语言生成示例

空间数据自然语言查询的可执行语言生成示例如表 A1 所示.

表 A1 可执行语言生成示例

自然语言查询 查询类型 转换后的可执行语言
Returns the distance between Mehringdamm
query distance (Mehringdamm, Alexanderplatz);
and Alexanderplatz.
What’s the direction from Mehringdamm to query direction (Mehringdamm, Alexanderplatz);
Alexanderplatz? 基础空间查询
How long is the border of Nanjing? query size (NJBorderLine);
Could you tell me the area of Nanjing query area (university feed filter [.Name = “Nanjing University”]
University? extract[GeoData]);
What are the roads that intersect the Zhixing query road feed filter [.GeoData intersects (road feed filter [.Name =
空间范围查询
Road? “Zhixing Road”] extract[GeoData])] consume;
Please tell me what are the 10 closest POIs query poi creatertree [GeoData] poi distancescan [NJBorderLine, 10]
最近邻居查询
to the Nanjing border line? consume;
What are the POIs in each district of query poi feed {a} district feed {b} symmjoin [.GeoData_a within
Nanjing? ..GeoData_b] consume;
What are the POIs within 1.5 kilometers query poi feed {a} district feed {b} symmjoin [distance (.GeoData_a,
from the center of each district in Nanjing? 空间Join查询 ..GeoData_b) <= 1500.0] consume;
Please find the POIs in each district that are query poi feed {a} university feed {b} symmjoin [distance
located within a 15-minute walk of a (.GeoData_a, ..GeoData_b) <= 1000.0] district feed {c} symmjoin
university. [.GeoData_a within ..GeoData_c] consume;