Re: [xml] Merging multiple XML trees

[Nic <nferrier tapsellferrier co uk>]

"Dustin D. Trammell" <dtrammell tippingpoint com> writes:

> Hello,
>
> First let me say that I am somewhat new to XML, and very new to libxml2.
> I tried searching this list's archives for previous discussion on what I
> am trying to do and also perused the reference section of functions on
> the xmlsoft.org reference area for functions that looked appropriate but
> so far have not found what I am looking for.  Also, please advise if
> this is not the appropriate place for this type of development question.
> Any pointers in the right direction would be most appreciated.
>
> What I am trying to do is have my application read in two XML trees.
> I'm doing this with the get_xml_doc_ptr() function.  The two XML trees
> have some common elements and some uncommon elements between them.  Many
> of the common elements have some common attributes and some uncommon
> attributes between the matching elements.  I want to merge these two
> trees into either a third, new tree or update one of the existing trees
> so that colliding elements, attributes, and values are preserved from
> one of the trees and all uncommon elements between the two trees will be
> available in the resultant tree.  For example:
>
> Tree A:
> <myelement attribute="yes!">
> 	<subelement1/>
> 	<subelement3/ threeattribute1="yes!"/>
> </myelement>
>
> Tree B:
> <myelement attribute="no!">
> 	<subelement1 oneattribute="yes!"/>
> 	<subelement2 twoattribute="yes!"/>
> 	<subelement3 threeattribute2="yes!"/>
> </myelement>
>
> Merging the two, and preserving values from Tree A, the resultant tree
> would look like the following.  Note the preservation of Tree A's
> colliding "attribute" attribute of the "myelement" element and the
> aggregation of all the subelements and subelement attributes in the
> subelement section:
>
> <myelement attribute="yes!">
> 	<sublement1 oneattribute="yes!"/>
> 	<subelement2 twoattribute="yes!"/>
> 	<subelement3 threeattribute1="yes!" threeattribute2="yes!"/>
> </myelement>
>
> Does libxml2 provide any functions that already do this or some similar
> tree merger?  It would actually be more beneficial to merge elements
> individually, not the entire tree.  For example, consider the
> "myelement" elements above to not be the root nodes of the tree and both
> be somewhere deep within two other mostly unrelated XML documents.
> Seeking to both of them with something like XPath and then creating a
> new XML document out of just those element sections is really what I'm
> trying to accomplish.
>
> So far, from looking at the xmlsoft.org reference area's function lists
> it appears I may have to code up this functionality manually, building a
> third tree from scratch (not hard, I already found the tutorial on that
> one...) and then stepping through the two (or more) source trees element
> by element, attribute by attribute and cross checking them all manually
> before deciding what to add to the resultant third tree.  That seems
> fairly messy, and I'd rather not do that if the capability is already
> within libxml2 to do the bulk of the work itself.
>
> Thanks in advance for any advice or pointers in the right direction any
> of you may be able to provide...

I'd use xslt for this. It would be a lot easier to express.


Nic Ferrier