Vector Database Management

This page is a work in progress.
Please contribute if you have experience with anything that is still poorly documented.

Vector data processing

• See the GRASS vectorintro vector data processing help page.
• See also the GRASS databaseintro Database management help page.

• See also the Openoffice.org with SQL Databases wiki page

Database Support

AsciiText (.csv, etc.)

The GRASS 6 vector engine supports the following databases:

• v.in.ascii module help page
• v.out.ascii module help page

If your .csv file contains "quoted" strings containing commas, you can use the csv_dequote.pl script to parse them into a less-problematic form ready for import with v.in.ascii.

DBF

In GRASS 6, DBF is the default (local) DB used for GRASS vector attributes. It is easy to use but with the simplicity comes limited features. Such limits are 10 chars per column name and no support for SQL calculations in SELECT statements.

• GRASS grass-dbf DBF driver help page

SQLite

SQLite is another local database format, but much more featureful than DBF. It basically combines the power of real SQL databases with the advantage of local data storage (no server needed). A nice tool to directly work in the SQLite database is SQLite Database Browser. In GRASS 7, SQLite is the default (local) DB used for GRASS vector attributes.

• GRASS grass-sqlite SQLite driver help page
• SQLite Homepage

MySQL

• GRASS grass-mysql MySQL driver page
• GRASS grass-mesql MSQL embedded driver page
• MySQL Homepage

PostgreSQL

• GRASS grass-pg PostgreSQL driver page
• PostgreSQL Homepage

FileMaker Pro

William Kyngesburye wrote on the grass-user mailing list:

I think the key to access FileMaker DBs from GRASS is that the FileMaker ODBC connector is for OSX's iODBC, not UnixODBC.

GRASS 6 has an iODBC configure option. It's the same --with- odbc-* options, it just tries iodbc if it can't find unixodbc.

Oracle

This has been reported to work. Try ODBC and search the mailing list archive. The connection is done via OGR.

ODBC

• External DB support via ODBC (e.g. FileMaker Pro)
• unixODBC is required to make it work
• to configure you could use the graphical frontend ODBCConfig to configure your ODBC connection.

Example-entry in ~/.odbc.ini for usage of ODBC with PostgreSQL

  [dbname]
  Description         = PostgreSQL database for my project
  Driver              = postgres
  Trace               = No
  TraceFile           =
  Database            = mydb
  Servername          = myserver
  UserName            = myusername
  Password            = mysecretpasswd
  Port                = 5432
  Protocol            = 9.0.3
  ReadOnly            = No
  RowVersioning       = No
  ShowSystemTables    = No
  ShowOidColumn       = No
  FakeOidIndex        = No
  ConnSettings        =

Additionall you need to define the libraries to use for the different drivers in /etc/odbcinst.ini.

  [postgres]
  Description     = ODBC for postgres
  Driver          = /usr/lib/unixODBC/libodbcpsql.so
  Setup           = /usr/lib/unixODBC/libodbcpsqlS.so
  FileUsage       = 1

• GRASS grass-odbc ODBC driver page

Concepts and jargon

• GRASS 6 Terminology
• Table
  • Table column
  • Table row
  • Vector map layer
  • Each vector file has a special data field named "cat" (derived originally from "category"), filled with integers, that serves to identify each vector object. The 'cat' field also serves as a "key field" that can link each vector object with a corresponding record in an attributes table of a database (NB: 'cat' values do NOT have to be unique for vector objects, but DO have to be unique in an attributes table, permitting both one-to-one and many-to-one relationships). The attributes table must contain a key field, filled with integers (only integers are permitted to serve as key fields in GRASS), that matches the values in the vector 'cat' field.
  • A vector can be linked with more than one attribute table, using LAYERS. By default, every vector file has a LAYER 1 with a 'cat' field, filled with integers to identify each vector object. Additional *LAYERS*, along with their associated 'cat' fields, can be created using v.category. Each LAYER has its own independent 'cat' field that can be used to link with a separate attributes table.
  • For example, a vector file of cities can have *LAYER 1* whose 'cat' field links the vector points with an attributes table of demographic data; it can also have a LAYER 2, with an associated 'cat' field (independent of the 'cat' field of LAYER 1), linked to a different attributes table of economic data. Values in the 'cat' fields of different LAYERS can be the same or different. The 'cat' values and linked attributes table can be queried independently for each LAYER. In the example, one can query the cities by population from the demographics attributes table linked with the 'cat' field of LAYER 1, or query the cities by household income from the economic attributes table linked with the same vector points through the 'cat' field of LAYER 2. _MichaelBarton - 12 Nov 2005_

Common tasks

Background info find in vectorintro Vector data processing in GRASS GIS.

• Connect a DB (db.connect, v.db.connect)
• Copy a table (db.copy)
• Copy selected columns from a table
• Create a new table (v.db.addtable)
• Create a new column (v.db.addcol)
• Extract data via SQL query (v.extract, db.select)
• Low level access to DB
• The db.execute module
• Populate a DB (v.db.update etc.)
  • The v.to.db module

More Help

• database module help pages
• vector module help pages
• GRASS sql query help page
• SQL reference links
• MySQL reserved words list