An overview of spatial indexes in the geodatabase
ArcGIS uses spatial indexes to quickly locate features in feature classes. Identifying a feature, selecting features by pointing or dragging a box, and panning and zooming all require ArcMap to use the spatial index to locate features.
Feature classes in some geodatabases use a system of grids as the spatial index. When you create an empty feature class in the Catalog tree or import data to create a new feature class, you can choose a default grid size or specify your own. If you are unfamiliar with creating spatial index grids, use the default value provided; creating a poorly defined grid will increase the time needed to complete a spatial search.
The spatial index on a feature class in a file or ArcSDE geodatabase can have up to three grids. The additional grids allow feature classes with features of very different sizes to be queried faster. The size of each grid in a file or ArcSDE geodatabase feature class must be at least three times the previous grid size. However, for most feature classes, only a single grid size is required. You can modify the grid size at any time.
Spatial index grids are only used for feature classes in ArcSDE geodatabases that use the following geometry storage and database management system:
- Binary storage in Oracle or Microsoft SQL Server
- ST_Geometry storage in Oracle or IBM DB2
ArcMap does not use spatial index grid sizes in IBM Informix or PostgreSQL databases, which use R-tree spatial indexes, nor are spatial index grid sizes used for feature classes that use Microsoft SQL Server spatial types or Oracle Spatial for geometry storage. Because other strategies are used to locate features in these geodatabases, you can ignore grid sizes. You also will not be able to use ArcCatalog to update the spatial indexes of feature classes that do not use spatial grid indexes.
For more information on spatial indexes in file, personal, and ArcSDE geodatabases, see Setting spatial indexes.