Update 05-boundary-functions.md
This commit is contained in:
@@ -112,7 +112,7 @@ geom | WKB geometry
|
||||
|
||||
Overwrite a point geometry with a boundary geometry that contains it in your table
|
||||
|
||||
```SQL
|
||||
```sql
|
||||
UPDATE tablename
|
||||
SET the_geom = OBS_GetBoundary(the_geom, 'us.census.tiger.block_group')
|
||||
```
|
||||
@@ -145,7 +145,7 @@ geometry_id | a string identifier of a geometry in the Boundaries
|
||||
|
||||
Write the US Census block group geoid that contains the point geometry for every row as a new column in your table.
|
||||
|
||||
```SQL
|
||||
```sql
|
||||
UPDATE tablename
|
||||
SET geometry_id = OBS_GetBoundaryId(the_geom, 'us.census.tiger.block_group')
|
||||
```
|
||||
@@ -178,7 +178,7 @@ geom | a WGS84 polygon geometry
|
||||
|
||||
Use a table of `geometry_id`s (e.g., geoid from the U.S. Census) to select the unique boundaries that they correspond to and insert into a table called, `overlapping_polygons`. This is a useful method for creating new choropleths of aggregate data.
|
||||
|
||||
```SQL
|
||||
```sql
|
||||
INSERT INTO overlapping_polygons (the_geom, geometry_id, point_count)
|
||||
SELECT
|
||||
OBS_GetBoundaryById(geometry_id, 'us.census.tiger.county') As the_geom,
|
||||
|
||||
Reference in New Issue
Block a user