A SQLAlchemy dialect for the FlexODBC driver
This dialect is mainly intended to offer an easy way to access the DataFlex flat-file databases of older or EOL'd application-specific softwares. It is designed for use with the ODBC driver available from FlexQuarters
-
The FlexODBC driver from FlexQuarters.
The library was written and tested against v3, but it should work with v4 as well.The library has been re-written from the ground up and tested to work against the latest available version of FlexODBC (4.0.27.2). -
32-bit Python. Neither of the available FlexODBC driver versions support or come in a 64-bit version.
This dialect requires SQLAlchemy and pyodbc. They are both specified as
requirements so pip will install them if they are not already in
place. To install separately, just:
pip install -U SQLAlchemy pyodbc
PyPI-published version:
pip install -U sqlalchemy-dataflex
Absolute bleeding-edge (probably buggy):
pip install -U git+https://github.com/the-wondersmith/sqlalchemy-dataflex
Create an ODBC DSN (Data Source Name) that points to the directory
containing your DataFlex FILELIST.cfg and .DAT table.
Tip: For best results, be sure to select the 4-digit date option and the
.option for decimals
Then, in your Python app, you can connect to the database via:
from sqlalchemy_dataflex import *
from sqlalchemy import create_engine
from sqlalchemy.orm import sessionmaker
db = create_engine("dataflex+pyodbc://@your_dsn", echo=False)
super_session = sessionmaker(bind=db)
super_session.configure(autoflush=True, autocommit=True, expire_on_commit=True)
session = super_session()NOTE: It is highly recommended that you use the datatype classes from the package itself in place of the ones that are usually imported from SQLAlchemy. The testing suite doesn't use the SQLAlchemy type classes, so any issues that might arise from their use with the dialect won't be caught by the tests.
SQLAlchemy-DataFlex is based on SQLAlchemy-access, which is part of the [SQLAlchemy Project] (https://www.sqlalchemy.org) and tries to adhere to the same standards and conventions as the core project.
At the time of this writing, it's unlikely that SQLAlchemy-DataFlex actually does comply with the aforementioned standards and conventions. That will be rectified (if and when) in a future release.
SQLAlchemy maintains a Community Guide detailing guidelines on coding and participating in that project.
While I'm aware that this project could desperately use the participation of anyone else who actually knows what they're doing, DataFlex and the FlexODBC driver are so esoteric and obscure (at the time of this writing) that I don't reasonably expect anyone to actually want to. If I am mistaken in that belief, please God submit a pull request.
This library technically works, but it isn't feature-complete so to
speak. The FlexODBC driver only supports a very limited subset of SQL
commands and queries, and doesn't always respond to pyodbc's
get_info() inquiries the way that pyodbc is expecting. You can see
the complete list of the way it responds to all of the available pyodbc
get_info() inquiries here. If you are
interested in which features are currently lacking or absent in the dialect,
see the notes and comments littered throughout the testing suite here.
This library is derived almost in its entirety from the SQLAlchemy-Access library written by Gord Thompson. As such, and given that SQLAlchemy-access is distributed under the MIT license, this library is subject to the same licensure and grant of rights as its parent works SQLAlchemy and SQLAlchemy-Access.