How to determine Parcel Neighbors?

Question

i like to get the neighbors of the Parcels "P1" like following :

From B1 to B2 => P5

B2 => P4

From B2 to B3 => P3

From B3 to B4 => P2

From B4 to B1 => P6

i like to make this process automatic by a GIS software like ArcGIS or Postgis, i have thought to make a buffer of the parcel P1 , to get the lines of the boundary of P1 but i will not get the parcel P4 as neighbor .

any solution to get the neighbors like described above.

Answer 1

I have tried to make the possible solution simpler but it seems to be the only solution, but i like to have any suggestions .

• from polygons make lines (arcs) => lines.
• buffer polygons with 3m => polygon_buffer.
• make spatial join between polygon_buffer and lines with "one to many" , to get all the polygons touching the lines => sp_lines_buffer.
• for every polygon get its boundary lines, from "sp_lines_buffer" to get neighbor polygon for every line.

For Arc based adjacent it's enough , but for adding node based adjacent , we need to add those steps:

• from polygons get vertices => nodes.

• make spatial join between polygon_buffer and nodes with "one to many" , to get all the polygons touching the nodes => sp_nodes_buffer.

• take a node based neighbors only if the number of neighbors per node is > 3. if not we will ignore the node.

Answer 2

I would think using intersects would work. Let's say pid is your parcel ids

SELECT p1.pid, array_agg(b2.pid) As neighbors 
FROM parcels As p1 INNER JOIN parcels As b2 ON ST_Intersects(p1.geom, b2.geom)
WHERE p1.pid != b2.pid
GROUP BY p1.pid

If you would prefer a string instead of an array and are using PostgreSQL 9.0+, you can use

string_agg(b2.pid, ',') As neighbors

For lower versions of PostgreSQL

array_to_string(array_agg(b2.gid),',')

Answer 3

You do not mention the version of ArcGIS for Desktop that you are using but, since 10.1, there has been a Polygon Neighbors tool which I think would make this much easier to do:

Creates a table with statistics based on polygon contiguity (overlaps, coincident edges, or nodes).