The core functionality addadded by a spatial extension to a database is one or more ''spatial datatypes'', which allow for the storage of spatial data as attribute values in a table.<ref name="GISTBOK">{{cite web |last1=Yue |first1=P. |last2=Tan |first2=Z. |title=DM-03 - Relational DBMS and their Spatial Extensions |url=https://gistbok.ucgis.org/bok-topics/relational-dbms-and-their-spatial-extensions |website=GIS&T Body of Knowledge |publisher=UCGIS |access-date=5 January 2023}}</ref> Most commonly, a single spatial value would be a [[geometric primitive]] (point, line, polygon, etc.) based on the [[Data model (GIS)#Vector data model|vector data model]]. The datatypes in most spatial databases are based on the OGC [[Simple Features]] specification for representing geometric primitives. Some spatial databases also support the storage of [[Raster graphics|raster data]]. Because all geographic locations must be specified according to a [[spatial reference system]], spatial databases must also allow for the tracking and transformation of coordinate systems. In many systems, when a spatial column is defined in a table, it also includes a choice of coordinate system, chosen from a list of available systems that is stored in a lookup table.
The second major functionality extension in a spatial database is the addition of spatial capabilities to the query language (e.g., [[SQL]]); these give the spatial database the same [[Spatial analysis | query, analysis, and manipulation operations]] that are available in traditional GIS software. In most relational database management systems, this functionality is implemented as a set of new functions that can be used in SQL SELECT statements. Several types of operations are specified by the [[Open Geospatial Consortium]] standard:
