SQL/SDA: A Query Language forSupporting Spatial Data Analysis and ItsWeb-Based ImplementationHui Lin and Bo HuangAbstractÐAn important trend of current GIS development is to provide easy and effective access to spatial analysis functionalities forsupporting decision making based on geo-referenced data. Within the framework of the ongoing SQL standards for spatial extensions,a spatial query language, called SQL/SDA, has been designed to meet such a requirement. Since the language needs to incorporatethe important derivation functions (e.g., map-overlay and feature-fusion) as well as the spatial relationship and metric functions, thefunctionality of the FROM clause in SQL is developed in addition to the SELECT and WHERE clauses. By restructuring theFROM clause via a subquery, SQL/SDA is well-adapted to the general spatial analysis procedures using current GIS packages. Suchan extended SQL, therefore, stretches the capabilities of the previous ones. The implementation of SQL/SDA on the Internetadopts a hybrid model, which takes advantage of the Web GIS design methods in both the client side and server side. The client side ofSQL/SDA, programmed in the Java language, provides a query interface by introducing visual constructs such as icons, listboxes, andcomboboxes to assist in the composition of queries, thereby enhancing the usability of the language. The server side of SQL/SDA,which is composed of a query processor and Spatial Database Engine (SDE), carries out query processing on spatial databases afterreceiving user requests. It was demonstrated that using the familiar SELECT-FROM-WHERE statement instead of a single ad hoccommand or procedural commands like macro language in some GIS packages, SQL/SDA offers users an efficient option to performcomplicated multistep spatial data analyses on the Internet.Index TermsÐGeographical Information System (GIS), spatial analysis, query languages, spatial database, SQL, Internet, client/server, visual interface, Java.æ1INTRODUCTIONGEOGRAPHICAL Information System (GIS) technology hasexperienced an astonishing growth in recent decadesand such development is, in many ways, closely related tothe spatial analytic capabilities of GIS [17], [34], [40], [41],[42]. In fact, the capabilities to handle and analyze spatialdata are usually seen as the key characteristic thatdistinguishes a GIS from other information or computer-aided design systems.Currently, the rapid growth of the Internet has led to anincreasing concern over the Web GIS [1], [20]. AlthoughWeb GIS has been employed for spatial data access,transmission, simple retrieval, and mapping, the access tospatial analysis functions such as map overlay, buffer, andfeature fusion that are vital for GIS-based decision-makingis still very limited [6], [36]. This is, in part, due to the factthat there is no effective interface to support the expressionof user requests dealing with these analysis functions. In theWorld Wide Web (WWW) client/server environment, theclient usually sends a request to the server, which processesthe request and returns the result to the client. The design ofa spatial query language can facilitate such a query processthat is crucial for a Web GIS.The need for a spatial query language has been identifiedin the GIS arena [16]. In the last decade, many authors havedesigned their extended database languages (e.g., [2], [14],[25], [26], [31], [37], [48]). For a detailed overview; see [14],[24]. These SQL-like languages (Spatially Extended SQLs)introduce spatial data types (e.g., point, line, polygon, andimage) and spatial operators (e.g., distance, direction,overlap, and contain), allowing users to inquire aboutspatial features, primarily in terms of spatial relationships(e.g., ªA CONTAINS Bº) and metric constraints (e.g.,ªDISTANCE(A, B) < 10º). The Spatial SQL [14] also addsgraphical capabilities for the presentation of a query result.On the other hand, query languages using algebraicapproach have also been proposed [21], [22]. Theselanguages, however, concentrate on the representationand management of spatial data rather than stepwiseanalysis of spatial data (see also [44]).GIS, as a tool for spatial analysis, has been used inmany different fields. Svensson and Huang [44] discussedthe required functionalities (i.e., reclassification, measure-ment, overlay, neighborhood, and statistics) for a spatialanalysis system. In many cases, most of these operationscan be performed by current GIS packages in certainways. But, the complexity of comprehensive GIS softwarepackages has been an obstacle to their widespread use byapplication specialists. To this end, Geo-SAL was devisedon top of SAL, a query language different from SQL andQUEL, to support spatial data analysis. Since the syntaxIEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, VOL. 13, NO. 4, JULY/AUGUST 2001 671. The authors are with the Joint Laboratory for GeoInformation Science, TheChinese University of Hong Kong, Shatin, NT, Hong Kong.E-mail: {huilin, bohuang}@cuhk.edu.hk.Manuscript received 12 July 1999; accepted 16 Aug. 1999; posted to DigitalLibrary 6 Apr. 2001.For information on obtaining reprints of this article, please send e-mail to:[email protected], and reference IEEECS Log Number 110128.1041-4347/01/$10.00 ß 2001 IEEEand semantics of SAL are complicated, they inevitablyinfluence Geo-SAL considerably.Herring et al. [23] discussed user-defined extensions toSQL by using a macro expander. Such macroextensionsallow users to tailor their environment flexibly by usingspatial operators to suit their operational scenarios.In addition to spatial query languages, attempts havebeen also made on using a procedural language forcartographic modeling and spatial data analysis. Tomlin'sMap Algebra [46], which is most suited for the raster datastructure, is a typical example. Different interface designsfor implementing such an algebra, ranging from command-line to advanced graphical interfaces, are also assessed in[5]. The principles underlying these interfaces are consid-ered applicable to the design of web-top interfaces. Ingeneral, Map Algebra can be seen as a macro language ifseveral commands in it are used together. The comparisonbetween spatial query language and macro language will bediscussed in Section 6.Recently, there are two major efforts for standardizingthe storage and management of spatial data: SQL3 multi-media specification (SQL/MM) [43] and Open GIS SimpleFeatures Specification for SQL (OpenGIS SQL)