ODBC Plugin for Intake
Open Database Connectivity (ODBC) provides a standardised interface to many data-base systems and other data sources capable of similar functionality. It passes an SQL text query to the backend via a driver and routes the resultant data back to the calling program.
ODBC can access data in several Data Source Names (DSNs), each of which constitutes a specific connection to a specific database system of a given type.
To install the intake-odbc plugin, execute the following command
conda install -c intake intake-odbc
In most cases, the end-users of ODBC data do not configure their own DSNs, merely reference an existing system-wide configuration. Each DSN will be referenced by a unique name, and in general have different requirements in terms of further required parameters. All of this may be defined in the Intake catalog, so that each catalog entry executes a query on some ODBC backend without the user having to know how this is achieved.
In order to set up the ODBC system on a machine, general guidance can
be found on the turbodbc pages (this is the library used by this
plugin): odbc confguration,
database configuration.
- Create a development environment with
conda create. Then install the dependencies:
conda install -c intake intake turbodbc pandas unixodbc
- Development installation:
git clone https://github.com/ContinuumIO/intake-odbc
cd intake-odbc
python setup.py develop --no-deps
-
Create a DB to connect to, if you do not have an existing ODBC configuration. The module
tests/util.pycontains functions to start and stop MS SQL and PostgreSQL docker images that might be useful. You will generally need a driver, appropriate for your system and the DB backend you wish to connect to. -
Set up odbc config, e.g., the two
.inifiles provided. The default location of these files is system wide at/usr/local/etc, but you can override this by specifying an environment variable before starting python, e.g.,
export ODBCSYSINI=/path/to/intake-odbc/examples