Abstract With the growing prevalence of data-driven decision-making, Text-to-SQL has emerged as a promising solution to lower the barrier to data access by translating natural language queries into executable SQL statements, thereby enhancing user interaction with databases. Despite notable progress driven by deep learning and large language models, significant challenges persist in handling complex queries. This paper presents a comprehensive review of the Text-to-SQL task, structured around two core stages: natural language understanding and natural language translation. Methods are categorized along the technical evolution trajectory into four types: rule-based, machine learning-based, pre-trained language model-based, and large language model-based approaches. Unlike previous surveys, which focus on specific techniques or partial aspects of Text-to-SQL, our work offers a two-stage analytical framework, highlights the impact of large models, and provides a comparative analysis of limitations and trade-offs. Through detailed examination of accuracy, generalization, expressiveness, and computational cost, this survey presents insights into the advantages and disadvantages of each paradigm. Furthermore, the paper summarizes key benchmark datasets and evaluation metrics, and discusses directions to improve the robustness, security, and effectiveness of the existing Text-to-SQL systems.

Spatial databases play a vital role in a number of applications ranging from geographic information systems to location-based services. Application tasks typically access underlying spatial data to answer queries. However, non-experts lack the expertise necessary for formulating spatial queries. To fill in this gap, we propose an effective framework that translatesnaturallanguage queries overspatialdata into executable database queries, called NALSpatial. The framework consists of two core phases: (i)natural language understandingand (ii)natural language translation. Phase (i) extracts key entity information, comprehends the query intent and determines the query type by employing natural language processing techniques and deep learning algorithms. The key entities and query type are passed to phase (ii), which makes use of entity mapping rules and structured language models to construct executable database queries. NALSpatial supports dealing with five types of queries including (i)basic queries (e.g. distance and area), (ii)range queries, (iii)nearest neighbor queries, (iv)spatial join queriesand (v)aggregation queries. We develop NALSpatial in an open-source extensible database system SECONDO. Extensive experiments show that NALSpatial on average achieves response time of about 2.5 seconds, translatability of 95% and translation precision of 92%, outperforming three state-of-the-art methods.

Spatial databases play a vital role in many applications that access spatial data via appropriate queries. However, most application users lack the expertise necessary for formulating spatial queries. To fill in this gap, we propose an effective framework called NALSpatial that translates natural language queries over spatial data into executable database queries. NALSpatial consists of two core phases. The natural language understanding phase extracts key entity information, comprehends the query intent and determines the query type. The key entities and query type are passed to the subsequent natural language translation phase, which employs entity mapping rules and structured language models to construct executable database queries accordingly. We implement NALSpatial on the open-source extensible database system SECONDO to support range queries, nearest neighbor queries, spatial joins and aggregation queries. Extensive experiments show that NALSpatial on average achieves response time of about 2.5 seconds, translatability of 95% and translation precision of 92%, outperforming state-of-the-art natural language transformation methods.

Spatial databases have a wide range of applications such as urban planning, engineering management and data visualization for epidemic investigation. The number of users in spatial databases becomes significantly large due to the increasing demand of application requirements. Users send their queries and analysis tasks to the system and receive the corresponding feedback. However, there is a lack of research on natural language interfaces in spatial databases. In this demo, we present NALSD, a natural language transformation system designed specifically for spatial data queries. NALSD comprises two core components: (i) natural language understanding and (ii) natural language translation. The system enables automatic translation of natural language query on spatial data into executable language for the underlying database, and supports range query, nearest neighbor query and spatial join query. We demonstrate how to obtain database executable language and visualize query results based on the SECONDO system.

Moving objects databases (MODs) have been extensively studied due to their wide variety of applications including traffic management, tourist service and mobile commerce. However, queries in natural languages are still not supported in MODs. Since most users are not familiar with structured query languages, it is essentially important to bridge the gap between natural languages and the underlying MODs system commands. Motivated by this, we design a natural language interface for moving objects, named NALMO. In general, we use semantic parsing in combination with a location knowledge base and domain-specific rules to interpret natural language queries. We design a corpus of moving objects queries for model training, which is later used to determine the query type. Extracted entities from parsing are mapped through deterministic rules to perform query composition. NALMO is able to well translate moving objects queries into structured (executable) languages. We support four kinds of queries including time interval queries, range queries, nearest neighbor queries and trajectory similarity queries. We develop the system in a prototype system SECONDO and evaluate our approach using 240 natural language queries extracted from popular conference and journal papers in the domain of moving objects. Experimental results show that (i) NALMO achieves accuracy and precision 98.1 and 88.1, respectively, and (ii) the average time cost of translating a query is 1.47s.

In recent years, natural language processing has received increasing attention in the field of database. The main task is to bridge the gap between non-expert users who are not familiar with formal query languages and the system executable language. In this demo, we develop a tool called NLMO that performs the translation of NL questions to executable query languages over moving objects databases. The tool well supports spatio-temporal range queries and nearest neighbor queries. We demonstrate the tool by reporting executable languages translated from the natural language input.