XSLT Newbie Questions

[charlesw7]

Couple of newbie questions and one specific to the issue I am trying to solve.

In some cases, my XSLT will specify <xsl:output method="xml"/> but the output will be plain text with no XML tags or delimiters.

Second, I'm having some issues understanding templates. The basic format seems straightforward enough when I look at examples. However, I run into some issues trying to understand when the templates are evaluated, when to use an "apply-templates" after a match, and when to use a match vs. select.

Finally my specific issue.

My input XML may look something like this:
<ServiceTags>
<Service parm="123455">
<name>Joe Blow</name>
<address>1234 Oak Street</address>
</Service>
<Service parm="98989">
<name>Tom Swift</name>
<address>1721 3rd Street</address>
</Service>
<Service parm="123455">
<city>detroit </city>
<state>mi</state>
<Service parm="98989">
<city>san diego</city>
<state>ca</state>
</ServiceTags>

Note that the parm values are the same for a couple of elements, but the child elements are different. What I'm trying to do is to be able to treat the two sets of child elements as one. Any thoughts on this?

Thanks.

[jkmyoung]

Muenchian Grouping is one technique of grouping similar nodes together. (XSLT 1.0)
XSLT 2.0 has built-in functions for the grouping of nodes.

Muenchian grouping is like so:

<xsl:key name="ServiceByParm" match="Service" use="@parm"/>
 
<xsl:for-each select="//Service[count(.|(key('ServiceByParm', @parm)[1])=1]">
   <!--  Node group:-->
  <xsl:for-each select="key('ServiceByParm', @parm)">
      <!-- All nodes in this group -->
  </xsl:for-each>
</xsl:for-each>
 

Yeah, the logic of templates is a little bit complicated, since it's not like most linear progams that people are used to.

The output is not really affected by the method in most cases.

[charlesw7]

Thanks for your reply. I decided last night to stop messing around and read the XSLT spec from W3C and that helped clear things up somewhat. I had also run across the Muenchian grouping solution, which is helping me solve the problem. So I appreciate you adding the code specific to my example.

As for the template processing thing, I'll probably still struggle with it for awhile, but should catch on soon enough.

Thanks again for answering a stupid question.

[charlesw7]

Late Friday afternoon is probably not the best time to post this, but...
In any case, the code that was posted here has been a great help. However, I think there are still some things I'm missing.

One thing I neglected in my original example here was how I need to grab pieces of data from each node once the data is 'grouped'.

A better representation of this data is like:
<ServiceTags>
<Service parm="123455">
<name>Joe Blow</name>
<address>1234 Oak Street</address>
<name>Sam Blow</name>
<address>123 Elm Street</address>
<name>Eli Blow</name>
<address>998 7th ave</address>
</Service>
<Service parm="98981">
<name>Tom Brady</name>
<address>1721 3rd Street</address>
<name>Tom Sawyer</name>
<address>9191 old oak way</address>
<name>Tom Greenbaum</name>
<address>42 blue ribbon drive</address>
</Service>
<Service parm="123455">
<city>detroit </city>
<state>mi</state>
</Service>
<Service parm="98989">
<city>san diego</city>
<state>ca</state>
</Service>
</ServiceTags>

Notice that the <Service> elements which include the <name> and <address> elements may repeat within the <Service> element.

The problem should be simple to solve, but I'm stuck. I am trying to tie the names and addresses to the city and states where the parm values match between <Service> elements, and output something a little more reasonable - something that might look more like for each set.
<Service parm="98981">
<name></name>
<address></address>
<city></city>
<state></state>
</Service>

etc.

Point is I think I have figured out how things get grouped and sorted, but am having trouble accessing the correct node once I've done that.

Thanks.

[jkmyoung]

Problem is the name is not 'really' grouped to the address.
You'll probably have to use an axis:

<xsl:for-each select="name">
...
<xsl:value-of select="following-sibling::address[1]"/>
...
</xsl:for-each>
 

Note: the [1] limits the address node only to the one directly following the name.

http://www.zvon.org/xxl/XSLTreferenc...axesIndex.html