1. Create a Network Topology¶
osm2pgrouting is a convenient tool, and its focus to work
on OpenStreetMap data. There are several cases where osm2pgrouting can’t be used. Some network data already comes with a network
topology that can be used with pgRouting out-of-the-box. Often network data is
stored in Shape file format (
.shp) and we can use PostGIS’
shp2pgsql converter to import the data into a PostgreSQL database.
But what to do then?
In this chapter you will learn how to create a basic Routing Network Topology from a network data that does not have a routing Topology create the minimum attributes needed the Routing Network Topology.
1.1. Load network data¶
At first we will load OpenStreetMap sample data with osm2pgsql.
CITY="DS_TZ" cd ~/Desktop/workshop cp ~/data/osm/$CITY.osm.bz2 . createdb -U user osm_data psql -U user -d osm_data -c "CREATE EXTENSION postgis;" psql -U user -d osm_data -c "CREATE EXTENSION pgrouting;" osm2pgsql -U user -c -d osm_data --latlong --cache 5 --cache-strategy sparse $CITY.osm.bz2
Let’s see which tables have been created:
psql -U user -d osm_data -c "\d"
The table containing the road network data has the name
It consists of large amount of attributes.
psql -U user -d osm_data -c "\d planet_osm_roads"
It is common that road network data provides at least the following information:
Road link ID (gid)
Road class (class_id)
Road link length (length)
Road name (name)
Road geometry (the_geom)
This allows to display the road network as a PostGIS layer in GIS software, for example in QGIS. Though it is not sufficient for routing, because it doesn’t contain network topology information.
The next steps will use the PostgreSQL command line tool.
psql -U user osm_data
1.2. Create a Routing Network Topology¶
Having your data imported into a PostgreSQL database might require one more step for pgRouting.
Make sure that your data provides a correct Routing Network Topology, which consists of information about source and target identifiers for each road link. The results above, show that the network topology does not have any source and target information.
Creation of the Routing Network Topology is necessary.
PostGIS topology is not suitable for Routing.
pgRouting provides a general way for creating the Routing Network Topology
targetidentifiers to each road link
It can logically “snap” nearby vertices within a certain tolerance by assigning the same identifier.
Creates a vertices table related to it.
Creates the basic indices.
pgr_createTopology('<table>', <tolerance>, '<geometry column>', '<gid>')
For additional information see pgr_createTopology.
First add source and target column, then run the
… and wait.
Depending on the network size this process may take from minutes to hours.
Progress indicator can be read with PostgreSQL NOTICE
It will also require enough memory (RAM or SWAP partition) to store temporary data.
The dimension of the tolerance parameter depends on your data projection. Usually it’s either “degrees” or “meters”. In our example the geometry data projection to determine the tolerance:
SELECT find_srid('public','planet_osm_roads','way'); find_srid ----------- 4326 (1 row)
Based on this result the tolerance will be 0.00001
-- Add "source" and "target" column ALTER TABLE planet_osm_roads ADD COLUMN "source" integer; ALTER TABLE planet_osm_roads ADD COLUMN "target" integer; -- Run topology function SELECT pgr_createTopology('planet_osm_roads', 0.00001, 'way', 'osm_id');
1.3. Verify the Routing Network Topology¶
To verify that there is a basic Routing Network Topology:
Also a new table containing the vertices information was created:
idis the vertex identifier
the_geomis the geometry considered for that particular vertex identifier.
planet_osm_roadscorrespond to an
Additional columns are for analyzing the topology.
Now we are ready for our first routing query with pgr_dijkstra
1.4. Analyze and Adjust the Routing Network Topology¶
Analyzing the topology with pgr_analyzeGraph:
SELECT pgr_analyzeGraph('planet_osm_roads', 0.000001, the_geom := 'way', id := 'osm_id');
Adjusting the topology is not an easy task:
Is an isolated segment an error in the data?
Is an isolated segment because its on the edge of the bounding box?
Do the potential gaps found near dead ends because the tolerance was too small?
Are the intersections real intersections and need to be nodded?
Are the intersections bridges or tunnels and do not need to be nodded?
Depending on the application some adjustments need to be made.
Some topology manipulation functions help to detect and fix some of the topological errors in the data.