SDO_SAM.SPATIAL_CLUSTERS

Format

SDO_SAM.SPATIAL_CLUSTERS(
     tablename      IN VARCHAR2,
     colname        IN VARCHAR2,
     max_clusters   IN NUMBER,
     allow_outliers IN VARCHAR2 DEFAULT 'TRUE',
     tablepartition IN VARCHAR2 DEFAULT NULL
     ) RETURN SDO_REGIONSET;

Description

Computes clusters using the existing R-tree index, and returns a set of SDO_REGION objects where the geometry column specifies the boundary of each cluster and the geometry_key value is set to null.

Parameters

tablename

Name of the data mining table.

colname

Name of the column in tablename that holds the location coordinates.

max_clusters

Maximum number of clusters to obtain.

allow_outliers

TRUE (the default) causes outlying values (isolated instances) to be included in the spatial clusters; FALSE causes outlying values not to be included in the spatial clusters. (TRUE accommodates all data and may result in larger clusters; FALSE may exclude some data and may result in smaller clusters.)

tablepartition

Name of the partition in tablename.

Usage Notes

The clusters are computed using the spatial R-tree index on tablename.

Examples

The following example clusters the locations in cities into at most three clusters, and includes outlying values in the clusters.

SELECT * FROM
  TABLE(sdo_sam.spatial_clusters('PROJ_CITIES', 'LOCATION', 3, 'TRUE'));