XPath: ancestor:: axis vs. use of ../../../ etc.

Greetings.

Looking for a little better understanding of the results I'm getting.

--------------------
My initial use of abbreviated XPath axes like simple "../../@name"
worked. (See # 1 down below)

But then exploring *going up one more level* (../) so I could 
then *explicitly name* the element did _not_ work:  
        "../../../span-fill/@name"  (See # 2 down below)

[ I guess I thought the XPath would be more readable if 
   I showed *by name* which element had the @name attribute
  vs. leaving it just at a long string of "../../../" ]


parent::
Further looking into the parent::node() attribute showed me that it 
behaved the same as the use of both ../../ and ../../../
(that's what it's supposed to do, I'm sure!).  (See #s 3 and 4 below)


axis::
Finally, the use of "ancestor::span-fill/@name" did just what I wanted: 
it got the right data, and shows 
the name of the element explicitly.   (See # 5 below)

-------------------------------------------
>> My basic question: what is going on that 
	A.	"../../@name" 
is so different from 
	B.	"../../../span-fill/@name"
?
>> Or put another way, 
  what is it that the ancestor:: axis does
 to help the correct nodeset selection, such that it works,
  vs. whatever nodeset being built up by "B." that 
  evidently does not (correctly?) test the span-fill/@name value?

	C.	"ancestor::span-fill/@name"


A. YES. Gets correct info (a single <span-value> element contents: 
   "(California Prima OLB)"
B. NO. Gets the contents of ALL the <span-value> elements: 
  "(California Prima OLB California Prima NO OLB California SAVING OLB
etc. etc. )"
C. YES. Works correctly, like A.

========================
XML
========================
<span-fills>
   <span-fill name="checking-product">
      <segment-permutation name="1ca_2olb">
         <span-value>California Prima OLB</span-value>
      </segment-permutation>
      <segment-permutation name="1ca_2nonolb">
         <span-value>California Prima NO OLB</span-value>
      </segment-permutation>
      ...
   <span-fill name="saving-product">
      <segment-permutation name="1ca_2olb">
         <span-value>California SAVING OLB</span-value>
      </segment-permutation>      
       ...
   </span-fill>
   ...
<span-fills>

========================
XSLT
========================
<xsl:for-each 
select="document($basedir_span-fillLookupFilenameExt)//span-fills/span-f
ill"> 

<xsl:apply-templates 
select="segment-permutation/span-value[ancestor::span-fill/@name=$span-i
d]
 
[ancestor::segment-permutation/@name=$segment]/node()"/>


Thanks very much,
William Reilly
wreilly@xxxxxxxxxxx
Boston, Massachusetts, U.S.A.

===================================
1. "../../"
YES (1st approach)
<xsl:apply-templates
select="segment-permutation/span-value[../../@name=$span-id][../@name=$s
egment]"/> 


2. "../../../span-fill/"
NO  (2nd approach)
<xsl:apply-templates
select="segment-permutation/span-value[../../../span-fill/@name=$span-id
][../../segment-permutation/@name=$segment]"/>


3. "parent::node()/parent::node()/"
YES  (emulating 1st approach, but with 'parent::' axis)
<xsl:apply-templates
select="segment-permutation/span-value[parent::node()/parent::node()/@na
me=$span-id][parent::node()/segment-permutation/@name=$segment]"/>


4. "parent::node()/parent::node()/parent::node()/span-fill/"
NO   (emulating 2nd approach, but with 'parent::' axis)
<xsl:apply-templates
select="segment-permutation/span-value[parent::node()/parent::node()/par
ent::node()/span-fill/@name=$span-id][parent::node()/parent::node()/segm
ent-permutation/@name=$segment]"/>


5. "ancestor::span-fill/" 
YES  (arguably emulating 2nd approach, but with 'ancestor::' axis)
<xsl:apply-templates
select="segment-permutation/span-value[ancestor::span-fill/@name=$span-i
d][ancestor::segment-permutation/@name=$segment]"/>

+++ /END
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list