Re: [Webware-devel] Re: MiddleKit dump

[Wayne L. <wa...@la...>]

Problem solved.

When switching to the new Samples.csv generated from my DB, I wanted to
make sure I could revert to my original Samples file if something went
wrong, so I renamed my Samples.csv file to Samples.orig.csv.  Of course,
as soon as I went looking through the code I see:

filenames = glob(os.path.join(self._filename, 'Sample*.csv'))

I guess that convention makes sense, so you can split your Samples file
across multiple files, I just was not aware of it.

But regardless, as soon as I moved that old Samples file out of there,
my problem was solved.

Thanks again - now I can feel confident in being able to easily upgrade
the db schema!

btw, I was not subscribed to webware-devel with this address, so this 
might appear as a duplicate to the list.

Wayne


Jason Hildebrand wrote:
> On Mon, 2003-03-10 at 00:35, wa...@la... wrote:
> 
> 
>>I updated and gave this a try.  The new Samples.csv file looks great - and
>>I see the added Id column in the data.  However, when I run generate to
>>get new sql files, the InsertSamples.sql file has duplicate data for every
>>table, one with the Id field, and one without.
> 
> 
> Hi Wayne,
> 
> This is very strange -- I tested the code myself on two MySQL databases,
> and it worked fine for me.  Are you using MySQL or something else?  Are
> there duplicate samples for each and every table, or just some of them?
> Are you using inherited object models?  (I've never tried that.)
> 
> The dump code is really quite simple -- check out
> MiddleKit/Run/SQLObjectStore.py, and search for dumpObjectStore.  The
> code simply loops over all classes in the model, so if you're getting
> duplicates the tables must be in there twice (which I've certainly never
> seen happen).  I've cc'd the dev list -- maybe someone there has some
> insight.  Failing that, try poking around a bit and see if you can
> figure it out.
> 
> 
>>  This leads to the insert
>>failing, since the second insert is trying to insert data with a primary
>>key that already
>>exists.
>>
>>Perhaps a file has yet to be committed?
> 
> 
> Nope - I just double-checked.  Everything is in.
>