Thanks Jim.  You are right, I am looking for a workaround for a wedged
system.  I think the approach you suggest will work just fine, provided I
can find a copy of zope.xmlpickle.  When I went to check it out from SVN,
it had been "moved to a satellite as a preliminary to eggification" and my 
svn client dies on checkout with a "no repository found" message.

Is this a cockpit error on my part?

   svn co svn:// zope.zmlpickle

On Sun, 6 May 2007, Jim Fulton wrote:

> On May 5, 2007, at 4:48 PM, Dennis Allison wrote:
> >
> > Has anyone written a utility that copies out the content of a ZODB  
> > file
> > (in my case, a Zope Data.fs) and stores it in the local file system in
> > some rational way?
> >
> > Zope does not solve the problem because the backing Relational DBMS is
> > unavailable and the ZODB and the RDBMS are closely tied.
> Hm, so I take that to mean that you can't start Zope or load the  
> objects in ZODB (in Zope or even from a Python prompt).
> I suggest you use zope.xmlpickle, 
> trunk/, along with the file storage iterator or the file-storage  
> record iterator, which only iterates over current versions, to  
> convert the database records to XML.  The XML pickler tries to  
> produce usable, for some definition of usable, XML without knowing  
> anything about the data.  A good XML hacker should be able to convert  
> the XML to some more directly usable format.  We've had good success  
> using xml pickles for a couple of projects in recent years.  For  
> example, xml pickles were used to export the Zope 3 collector for  
> import to Launchpad.
> I really should write some utilities that put the iterators and xml  
> pickler together, although it shouldn't be hard.
> Jim
> --
> Jim Fulton                    mailto:[EMAIL PROTECTED]                Python 
> Powered!
> CTO                           (540) 361-1714                  
> Zope Corporation   


For more information about ZODB, see the ZODB Wiki:

ZODB-Dev mailing list  -

Reply via email to