David,

I don't know if it meets your needs of a .NET-based way of talking to MS 
Access, but you
might look into ResolverOne. See more below. I've also included some 
information on IronClad
which is the port of the python c-extensions to IronPython. MS is not the only 
commercial
concern supporting IronPython. I'm not advocating specifically for .NET, I'm 
just saying
maybe IronPython isn't dead yet.

Regarding IronPython and IronClad (the port of the C-extensions to IronPython) 
the IronClad 
home page http://code.google.com/p/ironclad/  says:

-----

Ironclad aims, in the long term, to allow IronPython users to transparently 
import and use any compiled CPython extensions. Ironclad works with IronPython 
2.6 and targets CPython 2.6 on 32-bit Windows; efforts to support other 
platforms are underway. 

The latest release is v2.6.0rc1, which should now be useful to quite a lot of 
people. The following packages have been confirmed to work (to a greater or 
lesser extent, as qualified parenthetically):
numpy 1.3.0 (over 1500 tests pass; Unicode data and numpy.distutils don't work).
numpy 1.4.0rc1 (over 1900 tests pass; similar issues to 1.3.0).
scipy 0.7.1 (over 2300 tests pass; a few parts are still rather slow)
bz2 from Python 2.6 (well tested)
_csv from Python 2.6 (a couple of issues remain)
Lots of other packages work reasonably well; try them and see. Regrettably, 
this release does not support PIL (IronPython bug) ormatplotlib (blame 
unclear). While h5py will run, it suffers from serious threading issues, and I 
can't recommend it for production use with IronPython. See the readme for more 
details.

Ironclad basically works by reimplementing the Python C API in C#, and 
performing a little bit of underhanded trickery to convince .pyd files to talk 
to our version of the API. The source distribution includes full tests and 
decent explanatory documentation.
Ironclad is fully open source, and is developed and supported by Resolver 
Systems and William Reade. He gratefully acknowledges Resolver Systems' 
generous and invaluable support, without which the project would never have got 
off the ground.
On that note, please download and play with Resolver One. It's a game-changing 
next-generation spreadsheet with full IronPython integration, it includes numpy 
support via Ironclad, and it is entirely a Good Thing.
 
-----

Resolver Systems produces a very affordable commercial spreadsheet called 
ResolverOne 
and has just released a cloud-based spreadsheet called Dirigible. ResolverOne 
has a web-based
interface option and can be used to talk to MS Access. For example, from this 
forum page:
http://www.resolversystems.com/forums/viewtopic.php?f=1&t=167&p=602&hilit=odbc&sid=7003f8c0d5916aa2dd56df0a8bdcb5e3#p602
 
by searching in the forums for ODBC:

Resolver One can make a database worksheet from any existing ODBC data source; 
if you want to create an ODBC data source pointing at an Access database, 
follow the instructions at http://support.microsoft.com/kb/305599  (note: those 
instructions specify Windows XP, but seem to me to be identical to the 
procedure for Vista).

ResolverOne is written in IronPython and also accepts user-coded formulas and 
other programming
in IronPython, and as far as I know, large parts of numpy and scipy are working 
now in ResolverOne.
The forum is very helpful so if you want to know if the product will do what 
you need you might
post your problem description there. There are also lots of programming 
examples here:
http://www.resolversystems.com/exchange/  and documentation here: 
http://www.resolversystems.com/documentation/index.php/Main_Page 
including some help pages on working with databases.

Enthought http://www.enthought.com/ is also working on a port of numpy/scipy 
but I don't know
anything more about it....

Best regards,

Melissa
-----
Dr. Melissa Rice, PhD
Full Moon Technical Solutions, LLC
14202 60th Ave, NW
Stanwood, WA 98292-4808
email: mailto:[email protected]
phone: 360-654-0709
cell: 425-923-7713




> On Fri, Oct 8, 2010 at 12:44 PM, David Goldsmith
<[email protected]>wrote:
> > > Date: Thu, 7 Oct 2010 16:46:11 -0700
> > > From: Adrian Klaver <[email protected]>
> > > Subject: Re: [SEAPY] Anyone here use ceODBC w/ MS Access DBs?
> > > To: Seattle Python Interest Group <[email protected]>
> > >
> > > On Thursday 07 October 2010 3:45:15 pm David Goldsmith wrote:
> > >> I'm having a helluva time:
> > >>
> > >> 0) Getting a simple INSERT INTO statement to work;
> > >
> > > I have not worked with particular module before but it follows the
> > > Python
> >
> > Db-api
> >
> > > so something like below should work where fld_1 is string
> > > field(char,varchar,text,etc) and fld_2 is an integer field.
> > >
> > > cur=conn.cursor()
> > > cur.execute("insert into table_name(fld_1,fld_2)
> > > values(?,?)",('test',1))
> > >
> > > For general view of dbapi:
> > > http://www.python.org/dev/peps/pep-0249/
> > >
> > >> 1) Finding a good, reasonably rich example set or "cookbook," esp. one
> > >> geared toward a DB newbie such as myself;
> > >
> > > Python cookbook?:
> > > http://code.activestate.com/recipes/langs/python/tags/database/
> >
> > Thanks Adrian.  Unfortunately, it's now moot: I've been informed that
> > I can't "touch" the DB w/ Python, 'cause it was hard enough convincing
> > our IT security people to let us touch it w/ MATLAB (they wanted only
> > .NET technology to touch it). :-(
> >
> > DG

Reply via email to