keys: repeated nodes from same key value

[Jeni Tennison <jeni dot tennison at epistemics dot co dot uk>]

Hi,

The question:  Should key() return repeated nodes from the same (string)
key value?

A bit of context:  I'm working on an application that takes interview
protocols and marks them up based on the definition of concepts/attributes
etc. in a 'knowledge model'.  The application takes partially marked-up
protocols and turns them into HTML containing a number of snippets like
(content changed to protect the innocent):

<p id="p-134" class="keng">
  Tell me about your <a href="chicken.html">chicken</a>
  <a href="farm.html">farm</a>.
</p>
<p id="p-135" class="expert">
  I run a <a href="chicken.html">chicken</a> <a href="farm.html">farm</a>.
  It had lots of <a href="chicken.html">chickens</a> on it until they
  escaped.
</p>

I then go over these protocols when generating pages for the concepts
themselves to give a table showing where in the protocols the concepts have
been mentioned.  To identify the snippets of interest, I use a key:

<xsl:key name="paras" match="html:*[@id][html:a]" use="html:a/@href" />

I then use key() to get these snippets:

<xsl:variable name="concept-id" select="@id" />
<xsl:for-each select="key('paras', concat($concept-id, '.html'))">
  ...
</xsl:for-each>

The problem:  When I get the node set using key() [in SAXON 5.4.1], I get
repeats of the snippets that contain more than one reference to the same
concept.  For example, if I use:

key('paras', 'chicken.html')

I get:

p-134
p-135
p-135

I kind of expected key() to produce a 'set' in the classical way: no
repeated nodes.  I expected the keys to be:

node		key name		key value
p-134		paras			chicken.html
p-134		paras			farm.html
p-135		paras			chicken.html
p-135		paras			farm.html

But it seems there's another key as well, a repeat of:

p-135		paras			chicken.html

Looking at the XSLT Recommendation, I can't see anything that explicitly
addresses this situation.  It does say (my emphasis):

<blockquote uri="http://www.w3.org/TR/xslt#key">
Thus, a node x has a key with name y and value z if **and only if** there
is an xsl:key element such that:
 * x matches the pattern specified in the match attribute of the xsl:key
element;
 * the value of the name attribute of the xsl:key element is equal to y; and
 * when the expression specified in the use attribute of the xsl:key
element is evaluated with x as the current node and with a node list
containing just x as the current node list resulting in an object u, then
either z is equal to the result of converting u to a string as if by a call
to the string function, or u is a node-set and z is equal to the
string-value of one **or more** of the nodes in u.
</blockquote>

The 'if and only if' and 'one or more' make me think there shouldn't be
repeated keys, but I'm loath to say that SAXON has a bug in it in case Mike
thinks I want a fight ;)  Does anyone else have an opinion?

Cheers,

Jeni

Dr Jeni Tennison
Epistemics Ltd, Strelley Hall, Nottingham, NG8 6PE
Telephone 0115 9061301 • Fax 0115 9061304 • Email
jeni.tennison@epistemics.co.uk



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