| For data is the kernel of GIS, the design and realization of spatial database is the precondition and key to the whole system design. To choose which kind of database for the storage and management of GIS data, how to realize the data management by using database system, how to make sure the integrality and consistency of geographic data and how to gain the high efficiency of the access to the spatial database are all the important questions that need to be solved during the whole system design.Through the analysis on the GIS data structure and the geographic data storage in several kinds of database system-RDBMS, OODBMS and object-relational DBMS (ORDBMS), the research concludes that the most convenient and efficient one is ORDBMS.At present, RDBMS is the widest used DBMS in the GIS. It can easily deal with the structured data by a series of table rows and lists, but there is much limitation in that of unstructured data.On the contrary, OODBMS is efficient and comprehensible in graphic data storage, but it don't support SQL based on industry standard, and neither have a compatible relationship with other kinds of DBMS. These greatly restrict its using area and foreground.ORDBMS brings object model into RDBMS and uses the intelligent of object model to deal with unstructured spatial data, so it avoids multifarious relational operation and not only remains the support for SQL standard, but also extends to the management of object data by SQL. It breaks through the limitation of pure OODBMS and the formula restrict of pure RDBMS.Oracle Spatial is the Oracle spatial groupware for fast data memory, access and analyze. This paper explores the ORACLE database' s capability of storing, managing, querying the spatial data, including other capability of supporting index of the spatial data, and transforming the data from one coordinate system to another coordinate system. |
