Given xml like the following:

<myXML>
  <ACTUAL_1_VALUE>
      <myTAG />
  </ACTUAL_1_VALUE>
  <ACTUAL_2_VALUE>
      <myTAG />
  </ACTUAL_2_VALUE>
etc...
</myXML>

What I would like to do is when I'm processing the myTAG in ACTUAL_2_VALUE, pull the name of the previous ACTUAL_# tag.

I've done some research online, and it looks like I probably want to use the preceding-sibling, but I'm not really sure how to do that since I'm actually wanting to pull the NAME of the PARENT's preceding-sibling, not the current tag's preceding sibling.

Any help is greatly appreciated. Thanks.

Here's what I'm tried, and got the following error:

<xsl:variable name="lastParent" select="preceding-sibling::name(..)" />

NodeTest expected here. preceding-sibling::-->name<--(..)


Thanks!

__________________
Thanks!

You want name(../preceding-sibling::*[1])

that is: go to the parent, then go to the first preceding sibling element, then get its name.

In XPath 2.0 you could also do

../preceding-sibling::*[1]/name()

if that seems more logical.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Thanks Michael, as always, you're a wealth of knowledge (and quick to respond)!

Thanks!

One follow up on this, what if I didn't want the first preceding sibling element, but instead wanted to select the closest preceding sibling that ended in a specific "VALUE"?

Thanks!

>the closest preceding sibling that ended in a specific "VALUE"

I read that as:

the closest preceding sibling whose name ends in a specific "VALUE"

It's not a good idea to build structural information into your element names - that's what attributes are for. But if you must, you can do

preceding-sibling::*[ends-with(local-name(), 'abcd')][1]

ends-with() is XPath 2.0.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Correct.

Agreed. Unfortunately, I don't create the XML, just process what I'm given, so I don't have the ability to make that change.

With my limited knowledge of 2.0, it sounds like in order to use it will involve some kind of software install/update, both of which can't be done for this project (besides, it would require every PC that wanted to view the XML correctly to perform the same upgrade/install, something that again may or may not be able to be done depending on each of their company's policies, so most likely not).

If it can't be done in 1.0 then I'll just have to report back that I'm unable to get it done with the current limitations, and see if they'll cough up the money to get the XML corrected so that the structural information is in the attributes instead of the element name, since that would be the less intrusive alternative.

Thanks again Michael, I appreciate the response, and feedback.

Thanks!