Is Python any good with MySQL?
paulb at infercor.no
Tue Nov 7 14:55:51 CET 2000
Keith MacDonald wrote:
> This is really the essence of the problem. Python has an enthusiastic
> following, who presumably know it inside out, but it just keeps throwing up
> hurdles at newcomers. There are various ways of judging success, but in
> this context, Python will only be really successful if new users can just
> pick it up and run with it. For most people, real life pressures preclude
> immersing yourself in each new technology that comes along. I'm prepared to
> devote time to learning Python on the job, as long as I'm learning to use it
> to solve my problems, not wasting time making it work at all.
I agree with these sentiments. Python was so attractive to me because it was
easy to get into and provided various tools in the standard set of modules which
made it easy to get certain jobs done. Perl supposedly has a very good database
framework whose closest equivalent for Python would be something like mxODBC.
Nevertheless, there are often too many problems just getting any modules to link
to the libraries provided by database systems.
Personally, I would recommend mxODBC on the basis of its strong DB-API support
and for the documentation that is available for it. Some self-promotion now
Still, it isn't easy to get ODBC or any other database connectivity working, but
I think that it is important to be able to document how it happened when it
suddenly does. As a result, people will then have a chance of starting a
configuration/installation activity with some hope that they will have a working
database module at the end.
I have played with a few different database systems and have had success with
Oracle (through the now deprecated oracledb module), Sybase ASE (through
ctsybase and through mxODBC and OpenLink tools), Sybase ASA (through mxODBC and
ASA's built-in ODBC support), and Solid (through mxODBC and OpenLink tools).
Some solutions that could be called "native", such as oracledb, seem to suffer
more from linking problems, but then this is supposedly an issue with Oracle
rather than with the approach taken by such solutions. ODBC solutions, at least
on UNIX, are going to require integration with other components, and this can be
time-consuming. Having a knowledge base of known successes could be rather
helpful to the new user.
I did try and get MySQL working but never had any real luck - it didn't help
that the MySQL site has (or had) a confusing download page, and after the hassle
of choosing the appropriate packages I couldn't get the required packages to
link and run reliably. Given that MySQL doesn't even support "proper"
SQL-oriented database system features, I decided that the reward in getting it
working would hardly be worth the effort expended. I now expect to be flamed by
satisfied MySQL users, but that only tells me that some people expect less from
such database systems than they ought to. ;-)
More information about the Python-list