19 Spatial Operators

SDO_ANYINTERACT

Format

SDO_ANYINTERACT(geometry1, geometry2);

Description

Checks if any geometries in a table have the ANYINTERACT topological relationship with a specified geometry. Equivalent to specifying the SDO_RELATE operator with 'mask=ANYINTERACT'.

See the section on the SDO_RELATE operator in this chapter for information about the operations performed by this operator and for usage requirements.

Keywords and Parameters

Returns

The expression SDO_ANYINTERACT(geometry1,geometry2) = 'TRUE' evaluates to TRUE for object pairs that have the ANYINTERACT topological relationship, and FALSE otherwise.

Usage Notes

See the Usage Notes for the SDO_RELATE operator in this chapter.

For an explanation of the topological relationships and the nine-intersection model used by Spatial, see Section 1.8.

Examples

The following example finds geometries that have the ANYINTERACT relationship with a query window (here, a rectangle with lower-left, upper-right coordinates 4,6, 8,8). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.)

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_ANYINTERACT(c.shape,
            SDO_GEOMETRY(2003, NULL, NULL,
              SDO_ELEM_INFO_ARRAY(1,1003,3),
              SDO_ORDINATE_ARRAY(4,6, 8,8))
            ) = 'TRUE';

    MKT_ID NAME       
---------- --------------------------------                             
         2 cola_b                                                               
         1 cola_a                                                               
         4 cola_d 

SDO_CONTAINS

Format

SDO_CONTAINS(geometry1, geometry2);

Description

Checks if any geometries in a table have the CONTAINS topological relationship with a specified geometry. Equivalent to specifying the SDO_RELATE operator with 'mask=CONTAINS'.

See the section on the SDO_RELATE operator in this chapter for information about the operations performed by this operator and for usage requirements.

Keywords and Parameters

Returns

The expression SDO_CONTAINS(geometry1,geometry2) = 'TRUE' evaluates to TRUE for object pairs that have the CONTAINS topological relationship, and FALSE otherwise.

Usage Notes

See the Usage Notes for the SDO_RELATE operator in this chapter.

For an explanation of the topological relationships and the nine-intersection model used by Spatial, see Section 1.8.

Examples

The following example finds geometries that have the CONTAINS relationship with a query window (here, a rectangle with lower-left, upper-right coordinates 2,2, 4,6). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.) In this example, only cola_a contains the query window geometry.

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_CONTAINS(c.shape,
            SDO_GEOMETRY(2003, NULL, NULL,
              SDO_ELEM_INFO_ARRAY(1,1003,3),
              SDO_ORDINATE_ARRAY(2,2, 4,6))
            ) = 'TRUE';

    MKT_ID NAME       
---------- --------------------------------                             
         1 cola_a

SDO_COVEREDBY

Format

SDO_COVEREDBY(geometry1, geometry2);

Description

Checks if any geometries in a table have the COVEREDBY topological relationship with a specified geometry. Equivalent to specifying the SDO_RELATE operator with 'mask=COVEREDBY'.

See the section on the SDO_RELATE operator in this chapter for information about the operations performed by this operator and for usage requirements.

Keywords and Parameters

Returns

The expression SDO_COVEREDBY(geometry1,geometry2) = 'TRUE' evaluates to TRUE for object pairs that have the COVEREDBY topological relationship, and FALSE otherwise.

Usage Notes

See the Usage Notes for the SDO_RELATE operator in this chapter.

For an explanation of the topological relationships and the nine-intersection model used by Spatial, see Section 1.8.

Examples

The following example finds geometries that have the COVEREDBY relationship with a query window (here, a rectangle with lower-left, upper-right coordinates 1,1, 5,8). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.) In this example, only cola_a is covered by the query window geometry.

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_COVEREDBY(c.shape,
            SDO_GEOMETRY(2003, NULL, NULL,
              SDO_ELEM_INFO_ARRAY(1,1003,3),
              SDO_ORDINATE_ARRAY(1,1, 5,8))
            ) = 'TRUE';

    MKT_ID NAME       
---------- --------------------------------                             
         1 cola_a

SDO_COVERS

Format

SDO_COVERS(geometry1, geometry2);

Description

Checks if any geometries in a table have the COVERS topological relationship with a specified geometry. Equivalent to specifying the SDO_RELATE operator with 'mask=COVERS'.

See the section on the SDO_RELATE operator in this chapter for information about the operations performed by this operator and for usage requirements.

Keywords and Parameters

Returns

The expression SDO_COVERS(geometry1,geometry2) = 'TRUE' evaluates to TRUE for object pairs that have the COVERS topological relationship, and FALSE otherwise.

Usage Notes

See the Usage Notes for the SDO_RELATE operator in this chapter.

For an explanation of the topological relationships and the nine-intersection model used by Spatial, see Section 1.8.

Examples

The following example finds geometries that have the COVERS relationship with a query window (here, a rectangle with lower-left, upper-right coordinates 1,1, 4,6). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.) In this example, only cola_a covers the query window geometry.

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_COVERS(c.shape,
            SDO_GEOMETRY(2003, NULL, NULL,
              SDO_ELEM_INFO_ARRAY(1,1003,3),
              SDO_ORDINATE_ARRAY(1,1, 4,6))
            ) = 'TRUE';

    MKT_ID NAME       
---------- --------------------------------                             
         1 cola_a

SDO_EQUAL

Format

SDO_EQUAL(geometry1, geometry2);

Description

Checks if any geometries in a table have the EQUAL topological relationship with a specified geometry. Equivalent to specifying the SDO_RELATE operator with 'mask=EQUAL'.

See the section on the SDO_RELATE operator in this chapter for information about the operations performed by this operator and for usage requirements.

Keywords and Parameters

Returns

The expression SDO_EQUAL(geometry1,geometry2) = 'TRUE' evaluates to TRUE for object pairs that have the EQUAL topological relationship, and FALSE otherwise.

Usage Notes

See the Usage Notes for the SDO_RELATE operator in this chapter.

For an explanation of the topological relationships and the nine-intersection model used by Spatial, see Section 1.8.

Examples

The following example finds geometries that have the EQUAL relationship with a query window (here, a rectangle with lower-left, upper-right coordinates 1,1, 5,7). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.) In this example, cola_a (and only cola_a) has the same boundary and interior as the query window geometry.

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_EQUAL(c.shape,
            SDO_GEOMETRY(2003, NULL, NULL,
              SDO_ELEM_INFO_ARRAY(1,1003,3),
              SDO_ORDINATE_ARRAY(1,1, 5,7))
            ) = 'TRUE';

    MKT_ID NAME       
---------- --------------------------------                             
         1 cola_a

SDO_FILTER

Format

SDO_FILTER(geometry1, geometry2, param);

Description

Uses the spatial index to identify either the set of spatial objects that are likely to interact spatially with a given object (such as an area of interest), or pairs of spatial objects that are likely to interact spatially. Objects interact spatially if they are not disjoint.

This operator performs only a primary filter operation. The secondary filtering operation, performed by the SDO_RELATE operator, can be used to determine with certainty if objects interact spatially.

Keywords and Parameters

Returns

The expression SDO_FILTER(geometry1,geometry2) = 'TRUE' evaluates to TRUE for object pairs that are non-disjoint, and FALSE otherwise.

Usage Notes

The SDO_FILTER operator must always be used in a WHERE clause and the condition that includes the operator should be an expression of the form SDO_FILTER(arg1, arg2) = 'TRUE'.

geometry2 can come from a table or be a transient SDO_GEOMETRY object, such as a bind variable or SDO_GEOMETRY constructor.

• If the geometry2 column is not spatially indexed, the operator indexes the query window in memory and performance is very good.

• If two or more geometries from geometry2 are passed to the operator, the ORDERED optimizer hint must be specified, and the table in geometry2 must be specified first in the FROM clause.

If geometry1 and geometry2 are based on different coordinate systems, geometry2 is temporarily transformed to the coordinate system of geometry1 for the operation to be performed, as described in Section 6.10.1.

In previous releases, the SDO_FILTER operator required a third parameter. Effective with Oracle Spatial release 10.1, the operator has only two parameters. For backward compatibility, any keywords for the third parameter that were supported in the previous release will still work; however, the use of those keywords is discouraged and is not supported for new uses of the operator.

Examples

The following example selects the geometries that are likely to interact with a query window (here, a rectangle with lower-left, upper-right coordinates 4,6, 8,8). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.)

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_FILTER(c.shape,
    SDO_GEOMETRY(2003, NULL, NULL,
      SDO_ELEM_INFO_ARRAY(1,1003,3),
      SDO_ORDINATE_ARRAY(4,6, 8,8))
    ) = 'TRUE';
 
    MKT_ID NAME                                                                 
---------- --------------------------------                                     
         2 cola_b                                                               
         1 cola_a                                                               
         4 cola_d 

The following example is the same as the preceding example, except that it includes only geometries where at least one side of the geometry's MBR is equal to or greater than 4.1. In this case, only cola_a and cola_b are returned, because their MBRs have at least one side with a length greater than or equal to 4.1. The circle cola_d is excluded, because its MBR is a square whose sides have a length of 4.

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_FILTER(c.shape,
    SDO_GEOMETRY(2003, NULL, NULL,
      SDO_ELEM_INFO_ARRAY(1,1003,3),
      SDO_ORDINATE_ARRAY(4,6, 8,8)),
    'min_resolution=4.1'
    ) = 'TRUE';
 
    MKT_ID NAME                                                                 
---------- --------------------------------                                     
         2 cola_b                                                               
         1 cola_a 

The following example selects the GID values from the POLYGONS table where the GEOMETRY column objects are likely to interact spatially with the GEOMETRY column object in the QUERY_POLYS table that has a GID value of 1.

SELECT A.gid 
  FROM Polygons A, query_polys B 
  WHERE B.gid = 1 
  AND SDO_FILTER(A.Geometry, B.Geometry) = 'TRUE';

The following example selects the GID values from the POLYGONS table where the GEOMETRY column object is likely to interact spatially with the geometry stored in the aGeom variable.

Select A.Gid
  FROM Polygons A 
  WHERE SDO_FILTER(A.Geometry, :aGeom) = 'TRUE';

The following example selects the GID values from the POLYGONS table where the GEOMETRY column object is likely to interact spatially with the specified rectangle having the lower-left coordinates (x1,y1) and the upper-right coordinates (x2, y2).

Select A.Gid 
  FROM Polygons A 
  WHERE SDO_FILTER(A.Geometry, sdo_geometry(2003,NULL,NULL,
                                   sdo_elem_info_array(1,1003,3),
                                   sdo_ordinate_array(x1,y1,x2,y2))
                   ) = 'TRUE';

The following example selects the GID values from the POLYGONS table where the GEOMETRY column object is likely to interact spatially with any GEOMETRY column object in the QUERY_POLYS table. In this example, the ORDERED optimizer hint is used and the QUERY_POLYS (geometry2) table is specified first in the FROM clause, because multiple geometries from geometry2 are involved (see the Usage Notes).

SELECT /*+ ORDERED */
  A.gid 
  FROM query_polys B, polygons A 
  WHERE SDO_FILTER(A.Geometry, B.Geometry) = 'TRUE';

Related Topics

• SDO_RELATE

SDO_INSIDE

Format

SDO_INSIDE(geometry1, geometry2);

Description

Checks if any geometries in a table have the INSIDE topological relationship with a specified geometry. Equivalent to specifying the SDO_RELATE operator with 'mask=INSIDE'.

See the section on the SDO_RELATE operator in this chapter for information about the operations performed by this operator and for usage requirements.

Keywords and Parameters

Returns

The expression SDO_INSIDE(geometry1,geometry2) = 'TRUE' evaluates to TRUE for object pairs that have the INSIDE topological relationship, and FALSE otherwise.

Usage Notes

See the Usage Notes for the SDO_RELATE operator in this chapter.

For an explanation of the topological relationships and the nine-intersection model used by Spatial, see Section 1.8.

Examples

The following example finds geometries that have the INSIDE relationship with a query window (here, a rectangle with lower-left, upper-right coordinates 5,6, 12,12). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.) In this example, only cola_d (the circle) is inside the query window geometry.

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_INSIDE(c.shape,
            SDO_GEOMETRY(2003, NULL, NULL,
              SDO_ELEM_INFO_ARRAY(1,1003,3),
              SDO_ORDINATE_ARRAY(5,6, 12,12))
            ) = 'TRUE';

    MKT_ID NAME       
---------- --------------------------------                             
         4 cola_d

SDO_JOIN

Format

SDO_JOIN(table_name1, column_name1, table_name2, column_name2, params, preserve_join_order) RETURN SDO_ROWIDSET;

Description

Performs a spatial join based on one or more topological relationships.

Keywords and Parameters

Returns

SDO_JOIN returns an object of SDO_ROWIDSET, which consists of a table of objects of SDO_ROWIDPAIR. Oracle Spatial defines the type SDO_ROWIDSET as:

CREATE TYPE sdo_rowidset as TABLE OF sdo_rowidpair;

Oracle Spatial defines the object type SDO_ROWIDPAIR as:

CREATE TYPE sdo_rowidpair AS OBJECT
   (rowid1  VARCHAR2(24),
    rowid2  VARCHAR2(24));

In the SDO_ROWIDPAIR definition, rowid1 refers to a rowid from table_name1, and rowid2 refers to a rowid from table_name2.

Usage Notes

SDO_JOIN is technically not an operator, but a table function. (For an explanation of table functions, see Oracle Database PL/SQL Language Reference.) However, it is presented in the chapter with Spatial operators because its usage is similar to that of the operators, and because it is not part of a package with other functions and procedures.

This table function is recommended when you need to perform full table joins.

The geometries in column_name1 and column_name2 must have the same SRID (coordinate system) value and the same number of dimensions.

For best performance, use the /*+ ORDERED */ optimizer hint, and specify the SDO_JOIN table function first in the FROM clause.

If a table is version-enabled (using the Workspace Manager feature), you must specify the <table_name>_LT table created by Workspace Manager. For example, if the COLA_MARKETS table is version-enabled and you want to perform a spatial join operation on that table, specify COLA_MARKETS_LT (not COLA_MARKETS) with the SDO_JOIN table function. (However, for all other Spatial functions, procedures, and operators, do not use the <table_name>_LT name.)

The SDO_JOIN table function is not supported on partitioned tables.

Table 19-3 shows the keywords for the params parameter.

Optimizing Self-Joins

If you are performing a self-join (that is, if table_name1 and table_name2 specify the same table), you can improve the performance by optimizing the self-join.

If SDO_JOIN is called without a mask (for example, ANYINTERACT) or distance specification, it compares only the index structure of the two geometry columns being joined. This can quickly identify geometry pairs that are "likely" to interact. If SDO_JOIN is called with a mask or distance specification, after the index is used to identify geometry pairs that are likely to interact, geometry coordinates are also compared to see if the geometry pairs actually do interact. Coordinate comparison is the most expensive part of the SDO_JOIN operation.

In a self-join, where the same geometry column is compared to itself, each geometry pair is returned twice in the result set. For example:

• For the geometry pair with ID values (1,2), the pair (2,1) is also returned. The undesired effect in SDO_JOIN is that the coordinates of the same geometry pair are compared twice, instead of once.

• ID pairs that are equal are returned twice. For example, a table with 50,000 rows will return ID pair (1,1) twice, ID pair (2,2) twice, and so on. This is also an undesired effect.

When calling SDO_JOIN in a self-join scenario, you can eliminate the undesired effects by eliminating duplicate comparison of geometry pairs and all coordinate comparisons where the ID values of the pairs match. This optimization uses SDO_JOIN for the primary filter only, and calls the SDO_GEOM.RELATE function to compare geometry coordinates. The following statement accomplishes this optimization by adding "AND b.rowid < c.rowid" as a predicate to the WHERE clause.

SQL> set autotrace trace explain
SQL> SELECT /*+ ordered use_nl (a,b) use_nl (a,c) */ b.id, c.id
           FROM TABLE(sdo_join('GEOD_STATES','GEOM','GEOD_STATES','GEOM')) a,
                       GEOD_STATES b,
                       GEOD_STATES c
          WHERE a.rowid1 = b.rowid
               AND a.rowid2 = c.rowid
               AND b.rowid < c.rowid
               AND SDO_GEOM.RELATE (b.geom, 'ANYINTERACT', c.geom, .05) = 'TRUE'
 
Execution Plan
----------------------------------------------------------
Plan hash value: 1412731386
---------------------------------------------------------------------------------------------------
| Id  | Operation                           | Name        | Rows  | Bytes | Cost (%CPU)| Time     |
---------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT                    |             |     1 |  1124 | 12787   (1)| 00:02:34 |
|   1 |  NESTED LOOPS                       |             |     1 |  1124 | 12787   (1)| 00:02:34 |
|   2 |   NESTED LOOPS                      |             |  4574 |  2514K|  8206   (1)| 00:01:39 |
|   3 |    COLLECTION ITERATOR PICKLER FETCH| SDO_JOIN    |       |       ||          |
|*  4 |    TABLE ACCESS BY USER ROWID       | GEOD_STATES |     1 |   561 |1   (0)| 00:00:01 |
|*  5 |   TABLE ACCESS BY USER ROWID        | GEOD_STATES |     1 |   561 |1   (0)| 00:00:01 |
Predicate Information (identified by operation id):
---------------------------------------------------
   4 - access(CHARTOROWID(VALUE(KOKBF$)))
   5 - access(CHARTOROWID(VALUE(KOKBF$)))
       filter("B".ROWID<"C".ROWID AND
               "SDO_GEOM"."RELATE"("B"."GEOM",'ANYINTERACT',"C"."GEOM",.05)='TRUE')
 
SQL> set autotrace off

In the preceding example, It is very important that AND b.rowid < c.rowid be before the call to SDO_GEOM.RELATE in the WHERE clause. This will omit the undesired scenarios for the invocation of the SDO_GEOM.RELATE function. Also, note that the example uses the ORDERED and USE_NL hints, and that the execution plan does not contain TABLE ACCESS FULL or HASH JOIN.

Examples

The following example joins the COLA_MARKETS table with itself to find, for each geometry, all other geometries that have any spatial interaction with it. (The example uses the definitions and data from Section 2.1.) In this example, rowid1 and rowid2 correspond to the names of the attributes in the SDO_ROWIDPAIR type definition. Note that in the output, cola_d (the circle in Figure 2-1) interacts only with itself, and not with any of the other geometries.

SELECT /*+ ordered */ a.name, b.name
  FROM TABLE(SDO_JOIN('COLA_MARKETS', 'SHAPE',
                      'COLA_MARKETS', 'SHAPE',
                      'mask=ANYINTERACT')) c,
       cola_markets a,
       cola_markets b
  WHERE c.rowid1 = a.rowid AND c.rowid2 = b.rowid
  ORDER BY a.name;

NAME                             NAME                                           
-------------------------------- --------------------------------               
cola_a                           cola_c                                         
cola_a                           cola_b                                         
cola_a                           cola_a                                         
cola_b                           cola_c                                         
cola_b                           cola_b                                         
cola_b                           cola_a                                         
cola_c                           cola_c                                         
cola_c                           cola_b                                         
cola_c                           cola_a                                         
cola_d                           cola_d                                         
 
10 rows selected.

Related Topics

• SDO_RELATE

SDO_NN

Format

SDO_NN(geometry1, geometry2, param [, number]);

Description

Uses the spatial index to identify the nearest neighbors for a geometry.

Keywords and Parameters

Table 19-4 lists the keywords for the param parameter.

Returns

This operator returns the sdo_num_res number of objects from geometry1 that are nearest to geometry2 in the query. In determining how near two geometry objects are, the shortest possible distance between any two points on the surface of each object is used.

Usage Notes

The operator is disabled if the table does not have a spatial index or if the number of dimensions for the query window does not match the number of dimensions specified when the index was created.

The operator must always be used in a WHERE clause, and the condition that includes the operator should be an expression of the form SDO_NN(arg1, arg2, '<some_parameter>') = 'TRUE'.

The operator can be used in two ways:

• If all geometries in the layer are candidates, use the sdo_num_res keyword to specify the number of geometries returned.

• If any geometries in the table might be nearer than the geometries specified in the WHERE clause, use the sdo_batch_size keyword and use the WHERE clause (including the ROWNUM pseudocolumn) to limit the number of geometries returned.

As an example of the sdo_batch_size keyword, assume that a RESTAURANTS table contains different types of restaurants, and you want to find the two nearest Italian restaurants to your hotel. The query might look like the following:

SELECT r.name FROM restaurants r WHERE 
   SDO_NN(r.geometry, :my_hotel, 'sdo_batch_size=10') = 'TRUE' 
   AND r.cuisine = 'Italian' AND ROWNUM <=2; 

In this example, the ROWNUM <=2 clause is necessary to limit the number of results returned to no more than 2 where CUISINE is Italian. However, if the sdo_batch_size keyword is not specified in this example, and if sdo_num_res=2 is specified instead of ROWNUM <=2, only the two nearest restaurants are considered, regardless of their CUISINE value; and if the CUISINE value of these two rows is not Italian, the query may return no rows.

The sdo_batch_size value can affect the performance of nearest neighbor queries. A good general guideline is to specify the number of candidate rows likely to satisfy the WHERE clause. Using the preceding example of a query for Italian restaurants, if approximately 20 percent of the restaurants nearest to the hotel are Italian and if you want 2 restaurants, an sdo_batch_size value of 10 will probably result in the best performance. On the other hand, if only approximately 5 percent of the restaurants nearest to the hotel are Italian and if you want 2 restaurants, an sdo_batch_size value of 40 would be better.

You can specify sdo_batch_size=0, which causes Spatial to calculate a batch size that is suitable for the result set size. However, the calculated batch size may not be optimal, and the calculation incurs some processing overhead; if you can determine a good sdo_batch_size value for a query, the performance will probably be better than if you specify sdo_batch_size=0.

If the sdo_batch_size keyword is specified, any sdo_num_res value is ignored. Do not specify both keywords.

Specify the number parameter only if you are using the SDO_NN_DISTANCE ancillary operator in the call to SDO_NN. See the information about the SDO_NN_DISTANCE operator in this chapter.

If two or more objects from geometry1 are an equal distance from geometry2, any of the objects can be returned on any call to the function. For example, if item_a, item_b, and item_c are nearest to and equally distant from geometry2, and if sdo_num_res=2, two of those three objects are returned, but they can be any two of the three.

If the SDO_NN operator uses a partitioned spatial index (see Section 5.1.3), the requested number of geometries is returned for each partition that contains candidate rows based on the query criteria. For example, if you request the 5 nearest restaurants to a point and the spatial index has 4 partitions, the operator returns up to 20 (5*4) geometries. In this case, you must use the ROWNUM pseudocolumn (here, WHERE ROWNUM <=5) to return the 5 nearest restaurants.

If geometry1 and geometry2 are based on different coordinate systems, geometry2 is temporarily transformed to the coordinate system of geometry1 for the operation to be performed, as described in Section 6.10.1.

SDO_NN is not supported for spatial joins.

In some situations the SDO_NN operator will not use the spatial index unless an optimizer hint forces the index to be used. This can occur when a query involves a join; and if the optimizer hint is not used in such situations, an internal error occurs. To prevent such errors, you should always specify an optimizer hint to use the spatial index with the SDO_NN operator, regardless of how simple or complex the query is. For example, the following excerpt from a query specifies to use the COLA_SPATIAL_IDX index that is defined on the COLA_MARKETS table:

SELECT /*+ INDEX(c cola_spatial_idx) */ 
  c.mkt_id, c.name, ... FROM cola_markets c, ...;

However, if the column predicate in the WHERE clause specifies any nonspatial column in the table for geometry1 that has an associated index, be sure that this index is not used by specifying the NO_INDEX hint for that index. For example, if there was an index named COLA_NAME_IDX defined on the NAME column, you would need to specify the hints in the preceding example as follows:

SELECT /*+ INDEX(c cola_spatial_idx) NO_INDEX(c cola_name_idx) */ 
  c.mkt_id, c.name, ... FROM cola_markets c, ...;

(Note, however, that there is no index named COLA_NAME_IDX in the example in Section 2.1.)

If you join two or more tables with the SDO_NN operator, specify the LEADING hint for the outer table and the INDEX hint for the inner table (the table with the spatial index), and specify the inner table last. For example:

SELECT  /*+ LEADING(b)  INDEX(a cola_spatial_idx) */ a.gid 
   FROM cola_qry  b,  cola_markets  a
      WHERE b.gid =1 AND
          SDO_NN(a.shape, b.shape, 'querytype=window sdo_num_res=1')='TRUE';

For detailed information about using optimizer hints, see Oracle Database Performance Tuning Guide.

Examples

The following example finds the two objects from the SHAPE column in the COLA_MARKETS table that are nearest to a specified point (10,7). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.)

SELECT /*+ INDEX(c cola_spatial_idx) */
 c.mkt_id, c.name  FROM cola_markets c  WHERE SDO_NN(c.shape,
   sdo_geometry(2001, NULL, sdo_point_type(10,7,NULL), NULL,
   NULL),  'sdo_num_res=2') = 'TRUE';

    MKT_ID NAME
---------- --------------------------------
         2 cola_b
         4 cola_d

The following example uses the sdo_batch_size keyword to find the two objects (ROWNUM <=2), with a NAME value less than 'cola_d', from the SHAPE column in the COLA_MARKETS table that are nearest to a specified point (10,7). The value of 3 for sdo_batch_size represents a best guess at the number of nearest geometries that need to be evaluated before the WHERE clause condition is satisfied. (The example uses the definitions and data from Section 2.1.)

SELECT /*+ INDEX(c cola_spatial_idx) */ c.mkt_id, c.name
   FROM cola_markets c  
   WHERE SDO_NN(c.shape,  sdo_geometry(2001, NULL,
      sdo_point_type(10,7,NULL), NULL,  NULL),
      'sdo_batch_size=3') = 'TRUE'
   AND c.name < 'cola_d' AND ROWNUM <= 2; 

    MKT_ID NAME                                                                 
---------- --------------------------------                                     
         2 cola_b                                                               
         3 cola_c

See also the more complex SDO_NN examples in Section C.3.

Related Topics

• SDO_NN_DISTANCE

SDO_NN_DISTANCE

Format

SDO_NN_DISTANCE(number);

Description

Returns the distance of an object returned by the SDO_NN operator. Valid only within a call to the SDO_NN operator.

Keywords and Parameters

Returns

This operator returns the distance of an object returned by the SDO_NN operator. In determining how near two geometry objects are, the shortest possible distance between any two points on the surface of each object is used.

Usage Notes

SDO_NN_DISTANCE is an ancillary operator to the SDO_NN operator. It returns the distance between the specified geometry and a nearest neighbor object. This distance is passed as ancillary data to the SDO_NN operator. (For an explanation of how operators can use ancillary data, see the section on ancillary data in the chapter on domain indexes in Oracle Database Data Cartridge Developer's Guide.)

You can choose any arbitrary number for the number parameter. The only requirement is that it must match the last parameter in the call to the SDO_NN operator.

Use a bind variable to store and operate on the distance value.

Examples

The following example finds the two objects from the SHAPE column in the COLA_MARKETS table that are nearest to a specified point (10,7), and it finds the distance between each object and the point. (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.)

SELECT   /*+ INDEX(c cola_spatial_idx) */ 
   c.mkt_id, c.name, SDO_NN_DISTANCE(1) dist
   FROM cola_markets c  
   WHERE SDO_NN(c.shape,  sdo_geometry(2001, NULL, 
      sdo_point_type(10,7,NULL), NULL,  NULL),
      'sdo_num_res=2', 1) = 'TRUE' ORDER BY dist; 

    MKT_ID NAME                                   DIST                          
---------- -------------------------------- ----------                          
         4 cola_d                           .828427125                          
         2 cola_b                           2.23606798

Note the following about this example:

• 1 is used as the number parameter for SDO_NN_DISTANCE, and 1 is also specified as the last parameter to SDO_NN (after 'sdo_num_res=2').

• The column alias dist holds the distance between the object and the point. (For geodetic data, the distance unit is meters; for non-geodetic data, the distance unit is the unit associated with the data.)

Related Topics

• SDO_NN

SDO_ON

Format

SDO_ON(geometry1, geometry2);

Description

Checks if any geometries in a table have the ON topological relationship with a specified geometry. Equivalent to specifying the SDO_RELATE operator with 'mask=ON'.

See the section on the SDO_RELATE operator in this chapter for information about the operations performed by this operator and for usage requirements.

Keywords and Parameters

Returns

The expression SDO_ON(geometry1,geometry2) = 'TRUE' evaluates to TRUE for object pairs that have the ON topological relationship, and FALSE otherwise.

Usage Notes

See the Usage Notes for the SDO_RELATE operator in this chapter.

For an explanation of the topological relationships and the nine-intersection model used by Spatial, see Section 1.8.

Examples

The following example finds geometries that have the ON relationship with a query window (here, a rectangle with lower-left, upper-right coordinates 4,6, 8,8). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.) This example returns no rows because there are no line string geometries in the SHAPE column.

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_ON(c.shape,
            SDO_GEOMETRY(2003, NULL, NULL,
              SDO_ELEM_INFO_ARRAY(1,1003,3),
              SDO_ORDINATE_ARRAY(4,6, 8,8))
            ) = 'TRUE';

no rows selected

SDO_OVERLAPBDYDISJOINT

Format

SDO_OVERLAPBDYDISJOINT(geometry1, geometry2);

Description

Checks if any geometries in a table have the OVERLAPBDYDISJOINT topological relationship with a specified geometry. Equivalent to specifying the SDO_RELATE operator with 'mask=OVERLAPBDYDISJOINT'.

See the section on the SDO_RELATE operator in this chapter for information about the operations performed by this operator and for usage requirements.

Keywords and Parameters

Returns

The expression SDO_OVERLAPBDYDISJOINT(geometry1,geometry2) = 'TRUE' evaluates to TRUE for object pairs that have the OVERLAPBDYDISJOINT topological relationship, and FALSE otherwise.

Usage Notes

See the Usage Notes for the SDO_RELATE operator in this chapter.

For an explanation of the topological relationships and the nine-intersection model used by Spatial, see Section 1.8.

Examples

The following example finds geometries that have the OVERLAPBDYDISJOINT relationship with a line string geometry (here, a horizontal line from 0,6 to 2,6). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.) In this example, only cola_a has the OVERLAPBDYDISJOINT relationship with the line string geometry.

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_OVERLAPBDYDISJOINT(c.shape,
            SDO_GEOMETRY(2002, NULL, NULL,
              SDO_ELEM_INFO_ARRAY(1,2,1),
              SDO_ORDINATE_ARRAY(0,6, 2,6))
            ) = 'TRUE';

    MKT_ID NAME       
---------- --------------------------------                             
         1 cola_a

SDO_OVERLAPBDYINTERSECT

Format

SDO_OVERLAPBDYINTERSECT(geometry1, geometry2);

Description

Checks if any geometries in a table have the OVERLAPBDYINTERSECT topological relationship with a specified geometry. Equivalent to specifying the SDO_RELATE operator with 'mask=OVERLAPBDYINTERSECT'.

See the section on the SDO_RELATE operator in this chapter for information about the operations performed by this operator and for usage requirements.

Keywords and Parameters

Returns

The expression SDO_OVERLAPBDYINTERSECT(geometry1,geometry2) = 'TRUE' evaluates to TRUE for object pairs that have the OVERLAPBDYINTERSECT topological relationship, and FALSE otherwise.

Usage Notes

See the Usage Notes for the SDO_RELATE operator in this chapter.

For an explanation of the topological relationships and the nine-intersection model used by Spatial, see Section 1.8.

Examples

The following example finds geometries that have the OVERLAPBDYINTERSECT relationship with a query window (here, a rectangle with lower-left, upper-right coordinates 4,6, 8,8). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.) In this example, cola_a, cola_b, and cola_d have the OVERLAPBDYINTERSECT relationship with the query window geometry.

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_OVERLAPBDYINTERSECT(c.shape,
            SDO_GEOMETRY(2003, NULL, NULL,
              SDO_ELEM_INFO_ARRAY(1,1003,3),
              SDO_ORDINATE_ARRAY(4,6, 8,8))
            ) = 'TRUE';

    MKT_ID NAME                                                                 
---------- --------------------------------                                     
         2 cola_b                                                               
         1 cola_a                                                               
         4 cola_d 

SDO_OVERLAPS

Format

SDO_OVERLAPS(geometry1, geometry2);

Description

Checks if any geometries in a table overlap (that is, have the OVERLAPBDYDISJOINT or OVERLAPBDYINTERSECT topological relationship with) a specified geometry. Equivalent to specifying the SDO_RELATE operator with 'mask=OVERLAPBDYDISJOINT+OVERLAPBDYINTERSECT'.

See the section on the SDO_RELATE operator in this chapter for information about the operations performed by this operator and for usage requirements.

Keywords and Parameters

Returns

The expression SDO_OVERLAPS(geometry1,geometry2) = 'TRUE' evaluates to TRUE for object pairs that have the OVERLAPBDYDISJOINT or OVERLAPBDYINTERSECT topological relationship, and FALSE otherwise.

Usage Notes

See the Usage Notes for the SDO_RELATE operator in this chapter.

For an explanation of the topological relationships and the nine-intersection model used by Spatial, see Section 1.8.

Examples

The following example finds geometries that overlap a query window (here, a rectangle with lower-left, upper-right coordinates 4,6, 8,8). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.) In this example, three of the geometries in the SHAPE column overlap the query window geometry.

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_OVERLAPS(c.shape,
            SDO_GEOMETRY(2003, NULL, NULL,
              SDO_ELEM_INFO_ARRAY(1,1003,3),
              SDO_ORDINATE_ARRAY(4,6, 8,8))
            ) = 'TRUE';

    MKT_ID NAME                                                                 
---------- --------------------------------                                     
         2 cola_b                                                               
         1 cola_a                                                               
         4 cola_d 

SDO_RELATE

Format

SDO_RELATE(geometry1, geometry2, param);

Description

Uses the spatial index to identify either the spatial objects that have a particular spatial interaction with a given object such as an area of interest, or pairs of spatial objects that have a particular spatial interaction.

This operator performs both primary and secondary filter operations.

Keywords and Parameters

Returns

The expression SDO_RELATE(geometry1,geometry2, 'mask = <some_mask_val>') = 'TRUE' evaluates to TRUE for object pairs that have the topological relationship specified by <some_mask_val>, and FALSE otherwise.

Usage Notes

The operator is disabled if the table does not have a spatial index or if the number of dimensions for the query window does not match the number of dimensions specified when the index was created.

The operator must always be used in a WHERE clause, and the condition that includes the operator should be an expression of the form SDO_RELATE(arg1, arg2, 'mask = <some_mask_val>') = 'TRUE'.

geometry2 can come from a table or be a transient SDO_GEOMETRY object, such as a bind variable or SDO_GEOMETRY constructor.

• If the geometry2 column is not spatially indexed, the operator indexes the query window in memory and performance is very good.

• If two or more geometries from geometry2 are passed to the operator, the ORDERED optimizer hint must be specified, and the table in geometry2 must be specified first in the FROM clause.

If geometry1 and geometry2 are based on different coordinate systems, geometry2 is temporarily transformed to the coordinate system of geometry1 for the operation to be performed, as described in Section 6.10.1.

Unlike with the SDO_GEOM.RELATE function, DISJOINT and DETERMINE masks are not allowed in the relationship mask with the SDO_RELATE operator. This is because SDO_RELATE uses the spatial index to find candidates that may interact, and the information to satisfy DISJOINT or DETERMINE is not present in the index.

Although multiple masks can be combined using the logical Boolean operator OR, for example, 'mask=touch+coveredby', better performance may result if the spatial query specifies each mask individually and uses the UNION ALL syntax to combine the results. This is due to internal optimizations that Spatial can apply under certain conditions when masks are specified singly rather than grouped within the same SDO_RELATE operator call. (There are two exceptions, inside+coveredby and contains+covers, where the combination performs better than the UNION ALL alternative.) For example, consider the following query using the logical Boolean operator OR to group multiple masks:

SELECT a.gid
  FROM polygons a, query_polys B
  WHERE B.gid = 1
  AND SDO_RELATE(A.Geometry, B.Geometry,
                   'mask=touch+coveredby') = 'TRUE';

The preceding query may result in better performance if it is expressed as follows, using UNION ALL to combine results of multiple SDO_RELATE operator calls, each with a single mask:

SELECT a.gid
      FROM polygons a, query_polys B
      WHERE B.gid = 1
      AND SDO_RELATE(A.Geometry, B.Geometry,
                   'mask=touch') = 'TRUE'
UNION ALL
SELECT a.gid
      FROM polygons a, query_polys B
      WHERE B.gid = 1
      AND SDO_RELATE(A.Geometry, B.Geometry,
                   'mask=coveredby') = 'TRUE';

The following considerations apply to relationships between lines and a multipoint geometry (points in a point cluster). Assume the example of a line and a multipoint geometry (for example, SDO_GTYPE = 2005) consisting of three points.

• If none of the points has any interaction with the line, the relationship between the line and the point cluster is DISJOINT.

• If one of the points is on the interior of the line and the other two points are disjoint, the relationship between the line and the point cluster is OVERLAPBDYDISJOINT.

• If one of the points is on the boundary of the line (that is, if it is on the start point or end point of the line) and the other two points are disjoint, the relationship between the line and the point cluster is TOUCH.

• If one of the points is on the boundary of the line (that is, if it is on the start point or end point of the line), another point is on the interior of the line, and the third point is disjoint, the relationship between the line and the point cluster is OVERLAPBDYDISJOINT (not OVERLAPBDYINTERSECT).

Examples

The following examples are similar to those for the SDO_FILTER operator; however, they identify a specific type of interaction (using the mask keyword), and they determine with certainty (not mere likelihood) if the spatial interaction occurs.

The following example selects the geometries that have any interaction with a query window (here, a rectangle with lower-left, upper-right coordinates 4,6, 8,8). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1.)

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_RELATE(c.shape,
    SDO_GEOMETRY(2003, NULL, NULL,
      SDO_ELEM_INFO_ARRAY(1,1003,3),
      SDO_ORDINATE_ARRAY(4,6, 8,8)),
    'mask=anyinteract') = 'TRUE';

    MKT_ID NAME       
---------- --------------------------------                                     
         2 cola_b                                                               
         1 cola_a                                                               
         4 cola_d 

The following example is the same as the preceding example, except that it includes only geometries where at least one side of the geometry's MBR is equal to or greater than 4.1. In this case, only cola_a and cola_b are returned, because their MBRs have at least one side with a length greater than or equal to 4.1. The circle cola_d is excluded, because its MBR is a square whose sides have a length of 4.

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_RELATE(c.shape,
    SDO_GEOMETRY(2003, NULL, NULL,
      SDO_ELEM_INFO_ARRAY(1,1003,3),
      SDO_ORDINATE_ARRAY(4,6, 8,8)),
    'mask=anyinteract min_resolution=4.1') = 'TRUE';

    MKT_ID NAME       
---------- --------------------------------                                     
         2 cola_b                                                               
         1 cola_a                                                               

The following example selects the GID values from the POLYGONS table where the GEOMETRY column objects have any spatial interaction with the GEOMETRY column object in the QUERY_POLYS table that has a GID value of 1.

SELECT A.gid 
  FROM Polygons A, query_polys B 
  WHERE B.gid = 1 
  AND SDO_RELATE(A.Geometry, B.Geometry, 
                   'mask=ANYINTERACT') = 'TRUE';

The following example selects the GID values from the POLYGONS table where a GEOMETRY column object has any spatial interaction with the geometry stored in the aGeom variable.

SELECT A.Gid 
  FROM Polygons A 
  WHERE SDO_RELATE(A.Geometry, :aGeom, 'mask=ANYINTERACT') = 'TRUE';

The following example selects the GID values from the POLYGONS table where a GEOMETRY column object has any spatial interaction with the specified rectangle having the lower-left coordinates (x1,y1) and the upper-right coordinates (x2, y2).

SELECT A.Gid 
  FROM Polygons A 
  WHERE SDO_RELATE(A.Geometry, sdo_geometry(2003,NULL,NULL,
                                   sdo_elem_info_array(1,1003,3),
                                   sdo_ordinate_array(x1,y1,x2,y2)),
                     'mask=ANYINTERACT') = 'TRUE'; 

The following example selects the GID values from the POLYGONS table where the GEOMETRY column object has any spatial interaction with any GEOMETRY column object in the QUERY_POLYS table. In this example, the ORDERED optimizer hint is used and QUERY_POLYS (geometry2) table is specified first in the FROM clause, because multiple geometries from geometry2 are involved (see the Usage Notes).

SELECT /*+ ORDERED */
  A.gid 
  FROM query_polys B, polygons A 
  WHERE SDO_RELATE(A.Geometry, B.Geometry, 'mask=ANYINTERACT') = 'TRUE';

Related Topics

• SDO_FILTER

• SDO_JOIN

• SDO_WITHIN_DISTANCE

• SDO_GEOM.RELATE function

SDO_TOUCH

Format

SDO_TOUCH(geometry1, geometry2);

Description

Checks if any geometries in a table have the TOUCH topological relationship with a specified geometry. Equivalent to specifying the SDO_RELATE operator with 'mask=TOUCH'.

See the section on the SDO_RELATE operator in this chapter for information about the operations performed by this operator and for usage requirements.

Keywords and Parameters

Returns

The expression SDO_TOUCH(geometry1,geometry2) = 'TRUE' evaluates to TRUE for object pairs that have the TOUCH topological relationship, and FALSE otherwise.

Usage Notes

See the Usage Notes for the SDO_RELATE operator in this chapter.

For an explanation of the topological relationships and the nine-intersection model used by Spatial, see Section 1.8.

Examples

The following example finds geometries that have the TOUCH relationship with a query window (here, a rectangle with lower-left, upper-right coordinates 1,1, 5,7). (The example uses the definitions and data in Section 2.1 and illustrated in Figure 2-1.) In this example, only cola_b has the TOUCH relationship with the query window geometry.

SELECT c.mkt_id, c.name
  FROM cola_markets c
  WHERE SDO_TOUCH(c.shape,
            SDO_GEOMETRY(2003, NULL, NULL,
              SDO_ELEM_INFO_ARRAY(1,1003,3),
              SDO_ORDINATE_ARRAY(1,1, 5,7))
            ) = 'TRUE';
  FROM cola_markets c

    MKT_ID NAME       
---------- --------------------------------                             
         2 cola_b

SDO_WITHIN_DISTANCE

Format

SDO_WITHIN_DISTANCE(geometry1, aGeom, params);

Description

Uses the spatial index to identify the set of spatial objects that are within some specified distance of a given object, such as an area of interest or point of interest.

Keywords and Parameters

Returns

The expression SDO_WITHIN_DISTANCE(arg1, arg2, arg3) = 'TRUE' evaluates to TRUE for object pairs that are within the specified distance, and FALSE otherwise.

Usage Notes

The distance between two extended objects (nonpoint objects such as lines and polygons) is defined as the minimum distance between these two objects. The distance between two adjacent polygons is zero.

The operator is disabled if the table does not have a spatial index or if the number of dimensions for the query window does not match the number of dimensions specified when the index was created.

The operator must always be used in a WHERE clause and the condition that includes the operator should be an expression of the form:

SDO_WITHIN_DISTANCE(arg1, arg2, 'distance = <some_dist_val>') = 'TRUE'

The geometry column must have a spatial index built on it. If the data is geodetic, the spatial index must be an R-tree index.

SDO_WITHIN_DISTANCE is not supported for spatial joins. See Section 5.2.1.3 for a discussion on how to perform a spatial join within-distance operation.

Examples

The following example selects the geometries that are within a distance of 10 from a query window (here, a rectangle with lower-left, upper-right coordinates 4,6, 8,8). (The example uses the definitions and data described in Section 2.1 and illustrated in Figure 2-1. In this case, all geometries shown in that figure are returned.)

SELECT c.name FROM cola_markets c WHERE SDO_WITHIN_DISTANCE(c.shape,
  SDO_GEOMETRY(2003, NULL, NULL, SDO_ELEM_INFO_ARRAY(1,1003,3),
    SDO_ORDINATE_ARRAY(4,6, 8,8)),
  'distance=10') = 'TRUE';
 
NAME                                                                            
--------------------------------                                                
cola_b                                                                          
cola_a                                                                          
cola_c                                                                          
cola_d

The following example is the same as the preceding example, except that it includes only geometries where at least one side of the geometry's MBR is equal to or greater than 4.1. In this case, only cola_a and cola_b are returned, because their MBRs have at least one side with a length greater than or equal to 4.1. The trapezoid cola_c is excluded, because its MBR has sides with lengths of 3 and 2; and the circle cola_d is excluded, because its MBR is a square whose sides have a length of 4.

SELECT c.name FROM cola_markets c WHERE SDO_WITHIN_DISTANCE(c.shape,
  SDO_GEOMETRY(2003, NULL, NULL, SDO_ELEM_INFO_ARRAY(1,1003,3),
    SDO_ORDINATE_ARRAY(4,6, 8,8)),
  'distance=10 min_resolution=4.1') = 'TRUE';
 
NAME                                                                            
--------------------------------                                                
cola_b                                                                          
cola_a 

The following example selects the GID values from the POLYGONS table where the GEOMETRY column object is within 10 distance units of the geometry stored in the aGeom variable.

SELECT A.GID    FROM POLYGONS A    WHERE      SDO_WITHIN_DISTANCE(A.Geometry, :aGeom, 'distance = 10') = 'TRUE';

The following example selects the GID values from the POLYGONS table where the GEOMETRY column object is within 10 distance units of the specified rectangle having the lower-left coordinates (x1,y1) and the upper-right coordinates (x2, y2).

SELECT A.GID    FROM POLYGONS A    WHERE      SDO_WITHIN_DISTANCE(A.Geometry, sdo_geometry(2003,NULL,NULL,                              sdo_elem_info_array(1,1003,3),                              sdo_ordinate_array(x1,y1,x2,y2)),                      'distance = 10') = 'TRUE';

The following example selects the GID values from the POLYGONS table where the GID value in the QUERY_POINTS table is 1 and a POLYGONS.GEOMETRY object is within 10 distance units of the QUERY_POINTS.GEOMETRY object.

SELECT A.GID    FROM POLYGONS A, Query_Points B    WHERE B.GID = 1 AND      SDO_WITHIN_DISTANCE(A.Geometry, B.Geometry, 'distance = 10') = 'TRUE';

See also the more complex SDO_WITHIN_DISTANCE examples in Section C.2.

Related Topics

• SDO_FILTER

• SDO_RELATE