[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: Hierarchy

• From: Newsha Makooi <Newsha@i...>
• To: "'xml-dev@l...'" <xml-dev@l...>
• Date: Tue, 11 Jul 2000 14:52:06 -0600

Thanks for the info.  I am a bit confused about the second step you mentioned.  Do you see the <path="x/y/z"> parsed out as

<folder name="x"/><folder name="y"/><folder name="z"/>

        -OR-

<folder name="x">
        <folder name="y">
                <folder name="z">
                </folder>
        </folder>
</folder>
??

Then, the grouping would happen, right?
------------
Also, I could not find any examples for Preceding-Sibling. Just an explanation on page 736 (XSLT, Programmer's Reference).  Where can I find more info on Preceding-Sibling/Following-Sibling?

PS
Just to clarify my objective, I am using an XML document which has a listing of files and their location (absolute, such as Root\Folder1\Folder2\Folder3).  Each file has its own Path.  The XML feed has these files all over without any order.  I am trying to use XSLT to build a new XML document which groups all files that belong to the same folder Path.  So, I'll need to build a new XML tree structure and 'stick' the incoming XML entries into their appropriate location, thereby building hierarchy into the resulting XML.

Also, I am using MSXML3.

Thanks.
Newsha

-----Original Message-----
From: Kay Michael [mailto:Michael.Kay@i...]
Sent: Friday, July 07, 2000 2:30 AM
To: 'Newsha Makooi'
Subject: RE: Hierarchy

This kind of grouping problem is never easy in XSLT, and this is one of the
more complex ones! I think I'd break it up into several passes, either using
a sequence of stylesheets or using the node-set() extension (create a result
tree fragment, then process it further). The first pass might sort by path,
as you've already done, the second might parse the path (so path="x/y/z"
becomes <folder name="x"><folder name="y"><folder name="z">, and the third
might do grouping of adjacent entries that belong to the same folder: this
is then a classic grouping which you can do as
 
<xsl:for-each
select="folder[not(@name=preceding-sibling::folder[1]/@name)]">
   <folder name="{@name}">
       <xsl:for-each select="self::* |
following-sibling::folder[@name=current/@name]">
          [folder details]
 
Then the next trick is for this template to be applied recursively to each
level of folder.
 
Hope these ideas help.
 
Mike
 
 
 -----Original Message-----
From: Newsha Makooi [mailto:Newsha@i...]
Sent: 07 July 2000 01:17
To: 'Kay Michael'
Subject: Hierarchy

Hi Michael,

Sorry I am writing to you directly, but I think I am running out of options
and I needed your guidance if you have the time.

I am dealing with the following XML document (ZDocList.xml).

In my XSL document (TOC.xsl), I sort the XML document in order to group all
the elements.  Next, I would like to build a hierarchy based on the
FolderPath for all of the document entries.  But the XSL that I have only
builds the hierarchy for that node!  I want all of the nodes to be put into
the hierarchy based on their FolderPath.

In essence I want to end up with something similar to this:

<TOC>
        <entry folderpath="Level1">
                <entry folderpath="Level2">
                        <entry folderpath="">actual document</entry>
                </entry>
                <entry folderpath="Level2">
                        <entry folderpath="">actual document</entry>
                        <entry folderpath="">actual document</entry>
                        <entry folderpath="">actual document</entry>
                </entry>
        </entry>
        <entry folderpath="Level1">
                <entry folderpath="Level2">
                        <entry folderpath="">actual document</entry>
                        <entry folderpath="">actual document</entry>
                        <entry folderpath="">actual document</entry>
                </entry>
        </entry>
</TOC>

How can I get the above from what I have so far?  Would you please direct
me? 

Please ignore this if it takes too much of your time.  I don't want to be a
bother.  Support@Wrox has not been able to help because it does not pertain
to 'source code' mistakes in a book.

Thanks for your time and assistance.

Regards,
Newsha 

• Prev by Date: Re: Granularity
• Next by Date: XML and Dr. Dobbs
• Previous by thread: RE: Granularity
• Next by thread: Microsoft's DISCO proposal and XML packaging
• Index(es):
  • Date
  • Thread