<div dir="ltr">Dear Pythonoholics:<br> I have just released adodbapi version 2.2. This version adds support for Iron Python, which is not important to this group, but it did make me clean up the code. All classes are now new-style classes, and string exceptions are gone. I found out why .rowcount was so useless (you have to look in two places, and I was ignoring the more important one) so it works much better now.<br>
I have added a .cvtString method so it is now easy to get the package to deliver your numeric fields as strings (if you prefer) rather than the default decimal.Decimal. The unit test (used as a code sample) for user-defined type convertion was wrong. It happened to test correctly in CPython, but broke in Iron, so I finally figured out how it was supposed to work. (You pass it the type_code as found in .description[i][1], and the convertion function.) The test/sample is now correct.<br>
There are two user-submitted bug fixes (as in version 2.1.2) so multiple result sets from cursor.nextset() work even when some are empty.<br> --<br>For those who have no idea what I am talking about...<br><blurb><br>
adodbapi is a fully dbapi 2.0 compliant package for accessing SQL data, written in pure python.<br> It runs in either C python or Iron Python (Windows only - it depends on COM).<br>
All data is read and/or written using appropriate native python data
types such as datetime.datetime (for dates) and decimal.Decimal (for
currency and numeric). <br>
You may change the type conversions if you wish.<br> Everything works -- commit/rollback, stored procedures, multiple dataset return, etc, etc.<br> Tested with Jet (MS Access), MS SQL server, and mySQL engines. Probably works with any other suitable ADO DB source.<br>
</blurb><br>If you don't use cvs, you can also get the new version at <a href="http://sourceforge.net/projects/adodbapi" target="_blank">http://sourceforge.net/projects/adodbapi</a> . Put the unzipped folder in your /lib/site-packages directory.<br>
--<br>Vernon Cole<br><br></div>