[XML-SIG] xmlpickle.py ?!

[Mike Olson]

"M.-A. Lemburg" wrote:
> 
> Mike Olson wrote:
> >
> > "M.-A. Lemburg" wrote:
> > >
> > > Mike Olson wrote:
> > > >
> > > >
> > > > No matter how you do it, XPath won't look like python.  The "." is not
> > > > valid in XPath.  Attributes are accessed as with the attribute:: axis
> > > > (or @).  Ex.
> > > >
> > > > <Employees Department="Development"/>
> > > >   <Employee name="Mike">
> > > >     <Adress>1234 S West Way</Address>
> > > >   </Employee>
> > > > </Employees>
> > > >
> > > > To get Mike
> > > > /Employees/Employee[@name="Mike"]
> > > > To get Mikes Address with employee mike as the context:
> > > > Address[0]
> > >
> > > Hmm, looks like it would be more useful to map object names
> > > to element names in this case... it doesn't really make sense
> > > to access information on a type basis, e.g.
> > > /dictionary/item[@key="Mike"]. But then, structure is given by
> > > type, not object name. Oh well :-/
> >
> > I agree
> >
> > > BTW, how would one access "Mike" in this XML file without reverting to
> > > positional indexing ?
> > >
> > > <dictionary>
> > >         <item>
> > >                 <key>Mike</key>
> > >                 <value><address>1234 Main Street</address></value>
> > >         </item>
> > > </dictionary>
> > >
> > > It seems that these XPath lookups have to be context senstive...
> >
> > Not always.  You can have relative or absolute paths.  You can also have
> > paths that don't care about position.  Here are three examples.
> >
> > #1  Assume you have an unknown context with the exact document use:
> > /dictionary/item/key[text()='Mike']
> > #2  Assume you have a context, lets say at item
> > key[text()='Mike']
> > #3  Assume there are many dictionary objects in your document, all
> > scattered at different levels and you want to get all of the keys names
> > mike
> > //dictionary/key[text()='Mike]
> 
> The last one looks like a great way to search an XML file. I suppose
> you can then use the looked up tag as context, right ? In that
> case finding the item containing "Mike" as key wouldn't be hard.

Yes, but it is not recommended.  Imagine a 100 MB file.  the "//" step
will look at every node.  Yikes!  If you know the path it is recommed
that you use it.

Mike

> 
> Yah, I really should go and read the specs...
> 
> Thanks,
> --
> Marc-Andre Lemburg
> ______________________________________________________________________
> Business:                                      http://www.lemburg.com/
> Python Pages:                           http://www.lemburg.com/python/

-- 
Mike Olson				 Principal Consultant
mike.olson@fourthought.com               (303)583-9900 x 102
Fourthought, Inc.                         http://Fourthought.com 
Software-engineering, knowledge-management, XML, CORBA, Linux, Python