GRASS Vector Layers
Grass documentation provides basic information about vector attribute management, categories of vector features and vector layers. 
The aim of this tutorial is to show how to create, manipulate and display different layers of a vector map using relational database.
An organization provides information about routes to be driven by drivers. The organization's database consists of
- roads (road network)
- routes driven on these roads
Simple approach can be taken to present the routes to drivers. For every route a vector map could be created and overlayed over the road network vector map, i.e.
$ d.vect map=roads $ d.vect map=route01 color=green width=2
Above idea has a major flaw. If route data change or new route is added then vector maps has to be regenerated or removed.
Assuming that route network and routes data are stored in relational database, the database tables can be linked to road network map as its layers. Such layer information could be utilized to display maps, i.e.
$ d.vect map=roads $ d.vect map=roads layer=2 where="route_id=1" color=green width=2
Relation database should contain two tables
- road network table
- driven routes table
Road network table has to have at least two columns
- road segment id usually provided with road network data by vendor
- vectory category required by Grass to identify vector features
Routes table shall consist of
- route identifier
- id of road segment driven on a route, which references road segment id in road network table
Routes table also has to provide vector category information. This can be realized with a SQL view, i.e.
create view route_rn as select rn.cat as cat, rn.id as id, r.route_id as route_id from route r left join road_network rn on r.rn_id = rn.id;
Above tables and view allow to provide road network and driven routes information. They can be more complicated depending on application but for purpose of this tutorial such minimal approach is being kept.
Creating Grass Vector Layers
Sample data can be downloaded using one of the links
The archive contains basic relational database and Grass data based on Spearfish data
- gvl.sqlite file is SQLite relational database file with tables and view described above
- vector map roads containing road network map data
To start processing sample data unpack the archive and start Grass within PERMANENT location
$ unzip gvl-0.1.zip $ cd gvl-0.1 $ grass PERMANENT
Vector map roads provided with sample data has one layer of categories and it is not connected to any database table. This can be checked with commands
$ v.db.connect -p map=roads ERROR: Database connection for map <roads> is not defined in DB file $ v.category option=report input=roads Layer: 1 type count min max point 0 0 0 line 825 1 825 boundary 0 0 0 centroid 0 0 0 area 0 0 0 all 825 1 825
To connect vector map with two layers of data, vector categories have to be populated into these layers using following commands
$ v.category --o input=roads layer=2 output=tmprn $ g.remove vect=roads $ g.rename vect=tmprn,roads $ v.category option=report input=roads Layer: 1 type count min max point 0 0 0 line 825 1 825 boundary 0 0 0 centroid 0 0 0 area 0 0 0 all 825 1 825 Layer: 2 type count min max point 0 0 0 line 825 1 825 boundary 0 0 0 centroid 0 0 0 area 0 0 0 all 825 1 825
Now, database tables have to be connected to Grass vector map using v.db.connect command
$ v.db.connect -o map=roads driver=sqlite database=gvl.sqlite table=road_network $ v.db.connect -o map=roads layer=2 driver=sqlite database=gvl.sqlite table=route_rn $ v.db.connect -p map=roads Vector map <roads> is connected by: layer <1> table <road_network> in database <gvl.sqlite> through driver <sqlite> with key <cat> layer <2> table <route_rn> in database <gvl.sqlite> through driver <sqlite> with key <cat>
Road network table is connected as first layer and route information as second layer.
Almost all Grass commands accept layer parameter, which determines layer to be used by a command. Below, some examples are provided.
To display road network map and routes number 1 and 4
$ d.mon x0 $ d.vect map=roads $ d.vect map=roads layer=2 color=green width=8 where='route_id=1' $ d.vect map=roads layer=2 color=red width=4 where='route_id=4'
Creating Separate Route Map
To create separate vector map of route number 4
$ v.extract -t input=roads layer=2 output=r4 where='route_id=4'
Please note -t parameter, which forces Grass to not copy database table, which can be very costly operation.
To display vector r4, layer parameter has to be used as routing information is still in second layer
$ d.mon x0 $ d.vect map=r4 layer=2
Appendix: Creating Tutorial's Road Network and Routes Data
Sample data for this tutorial were extracted from Spearfish data. As it was quite interesting excersise, detailed information describing the process is provided below.
Road Network Data
Spearfish data provides road network map in roads vector. It contains few hundred vector lines - road network segments. Data associated with this vector map are very simple and they have to be extended
- unique category id for every vector road segment needs to generated
- unique id has to be assigned to every road segment
Below, vector roads is created with every vector line having unique category
$ v.category --o in=roads out=tmpmap option=del $ v.category --o in=tmpmap out=roads option=add $ g.remove vect=tmpmap
Vector map roads is connected to DBF file, which is no longer required, it has to be removed using command
$ v.db.droptable -f roads
Road network table needs to be populated with categories and road segment identifiers. First, road network table has to be connected to vector map
$ v.db.connect -o map=roads driver=sqlite database=gvl.sqlite table=road_network
Now, it is possible to upload categories of vector map roads into road network table
$ v.to.db map=roads option=cat
Road network data usually is provided with some road segment identifiers. The identifiers are usually set by vendor of road network data but they are missing in case of Spearfish data. Therefore, for the purpose of this tutorial they will be generated from categories of vector map roads. It can be done easily with SQL query
update road_network set id = cat;
Finally, vector map roads is being disconnected from road network table
$ v.db.connect -d map=road
Routes data were created using d.what.vect command. Road network vector map roads has to be displayed
$ d.mon x0 $ d.vect map=roads
Now, it can be queried for data. Run
$ d.what.vect -t map=roads > route-01.txt
Using mouse pointer, roads segments belonging to first route, can be identified. When done, file route.txt can be processed with awk or other tool to extract road segment information, which can be easily uploaded to database.