|
|
 next
|
  Subject: RE: Creating an XML Difference File
Author: Ivan Pedruzzi
Date: 15 Oct 2003 03:56 PM
|
Jason,
Building a generic XML diffs using XSLT is not trivial.
If you like we can help you to build a specific transformation but you
should provide small example documents and the result you would like to
have.
You also may want to take a look to the following resources:
Microsoft .NET Framework (free open source)
http://apps.gotdotnet.com/xmltools/xmldiff/
Java Platform (free open source)
http://diffxml.sourceforge.net/
Commercial product (commercial)
http://www.deltaxml.com/
Ivan
> -----Original Message-----
> From: stylus-studio-xslt Listmanager
> [mailto:listmanager@edn.exln.com]
> Sent: Wednesday, October 15, 2003 3:07 PM
> To: Recipients of 'stylus-studio-xslt' suppressed
> Subject: Creating an XML Difference File
>
>
> From: "Jason Kempel"
>
> I am trying to use XSL to create an html difference file
> between 2 nearly identical XML files. The XML files use 1
> element as header information, with several - elements
> nested below it. The header element has a few attributes,
> and each - has a few attributes. I want to display the
> differences in the attributes of the header, then identify
> either added, removed, or changed by checking item
> attributes for changes or for new/removed items from one
> revision of an xml file to the next.
>
> I have heard about using keys and templates to do this, but I
> am hunting for an example to start from, or even a command
> line utility that might do this, that I can call from inside
> a visual basic 6 application using MSXML 4.0.
>
> This is my first post to the group, so go easy on me :)
>
> Jason
>
>
>
> To reply: mailto:stylus-studio-xslt.7082@edn.exln.com
> To start a new topic: mailto:stylus-studio-xslt@edn.exln.com
> To login: http://edn.exln.com/~SSDN
>
|
 top
|
  Subject: RE: Creating an XML Difference File
Author: Lee Humphries
Date: 16 Oct 2003 06:59 PM
|
Hi Jason,
As one of my first training exercises in XSLT I wrote exactly what you're asking for now. Unfortunately I don't have it now, otherwise I'd just upload it.
You will absolutely need to use templates. Though for a completely generic XML comparator keys are going to be largely useless.
As Ivan said this is not trivial. If the structure of the files you're comparing is reasonably fixed then you'll be better off coming up with a very specific comparator (in which you'll be able to use keys so it will run much faster).
If it has to be generic though, here's one way that you can achieve this process.
First, combine the two files into one putting them both inside a single root node. This just simplifies things further down the track. You can either create a simple XSLT to do this, using the document() function to load at least one of the files into a variable. Or a quick and dirty text file operation can do it as well, just remember to strip off any xml declarations and PI's first.
Second, work on templates to compare attributes. Remember that the order of attributes should be unimportant.
Third, work on templates to traverse the two document hierarchies in parallel. This means you'll need a recursive template that keeps track of where you are in the current hierarchy, by that I mean a delimited string of node names. Why? When you're at one point in the first file you can pass that delimited string to another template for it to recursively work its way through the second file to see if it can find the same location. You'll need to be able to deal with missing elements and correctly matching up repeating elements. Correctly matching repeating elements means that your templates have to dig into the contents of both nodes to see if they truly match and not just rely on them having the same name. If they don't completely line up it has to figure out whether it has found the best match anyway.
When you throw namespaces in, things get a lot more interesting.
This should keep you busy for at least 3 or 4 days if you already have a reasonable understanding of XSLT.
Regards,
Lee Humphries
|
|
|
|
