Group together nodes that have similar content. then return count of uniques.

Hi idesta,
> /Call/ObjectSurveys/ObjectSurvey/Results/Result[Reference]

if you change your XPath like this
/Call/ObjectSurveys/ObjectSurvey/Results/Result[not(Reference = preceding::Result/Reference)]
you will only get the Result elements that have a unique Reference
(always the first one is selected)

Cheers!

idesta,

to get the number of unique entries, you just need to count them
count(//Call/ObjectSurveys/ObjectSurvey/Results/Result[not(Reference = preceding::Result/Reference)])

if you want to iterate over them in a DOM structure, use selectNodes

var aNodes = document.selectNodes("//Call/ObjectSurveys/ObjectSurvey/Results/Result[not(Reference = preceding::Result/Reference)]/Reference");

cheers

Hi Gertone.

I will give that a try.
I thought I might have to use keys?
Thanks G., Ill get back to you ;)

Idesta Si

Keys is another option,
it might turn out faster
this one is pure XPath and comes in handy when you just want to select unique nodes
cheers

I have over a thousand nodes to iterate on.
I would expect to do around a thousand iterations a day however so this is not life threatening, but I have to put this on a productions server.
So far in this project the thinking has been trying to push all the cpu-hungry load server side with post processing DOM nodes using the javascript functions invoked from the client-side browser - the function I demonstrated above, I am unsure of the advisedness of this in retrospect, but the work is done.

Simply, In short: All I am trying to do is group some 'result'.text nodes together. So I can show and hide the groups of divs in a browser by clicking somewhere... you get the idea.
I'll handle the hiding and showing etc... But I need the groups returning

In Long: The output I am parsing has a load of redundant structure around the results I need. The structure wraps results attached to groups of the original questions and these questions are wrapped in a tag of type 'surveys' all indentified together however by the value 'reference'. Would this affect my choice of Keys or your method? I think so? the \\ instruction would be very server heavy I believe, but I am not well informed.

All of this aside...

the important bits are: <xsl:for-each select="$er"> where I define the sets of data I am returning...

should this be a:      <xsl:for-each select="$er"> <!-- where this is the array of groups returned when sorting by reference
and then:                <xsl:for-each select="$er">  <!-- Individual structures within each group.

</
</

or am I completely missing how simple it should be to tag up:

<div id="page(n)" style="padding:0px;overflow:hidden;display:block;">
<div style="position: absolute; top: 155px; right: 80px; left: 114px; width: 250px; height: 20px;" class="page(n)" id="SiteName">------ </div></div>

where each (n) is the index of the groups returned and dynamically print that excerpt above (n) number of times into the resulting file.

I hope it is just really easy and Im being thick, because I dont get it.
Thanks.

sorry the point of all that was:

var aNodes = document.selectNodes("//Call/ObjectSurveys/ObjectSurvey/Results/Result[not(Reference = preceding::Result/Reference)]/Reference");

I am unsure how to use this? I suspect I cant basically.
thanks

Please could you explain how to do the same thing with keys?
I need the data returning in groups defined by the refrence rather than the 1st node.
I with to iterate through the resulting data set.

Many thanks.
Simon

I would be happy to spilt this into several questions if needs be.

count(//Call/ObjectSurveys/ObjectSurvey/Results/Result[not(Reference = preceding::Result/Reference)]) always returns 1?

is this the right direction?

      <xsl:key name="pages" match="/Call/ObjectSurveys/ObjectSurvey/Results/Result[Reference]" use="/Call/ObjectSurveys/ObjectSurvey/Results/Result[not(Reference = preceding::Result/Reference)]/Reference"/>

<xsl:template match="pages">
<xsl:apply-templates
      select="*[Reference and
                generate-id(.)=generate-id(key('pages', Reference))]" />
</xsl:template>


<xsl:template match="*[./Reference]">
      <!-- first column is the value of the partNum attribute -->
  <xsl:value-of select="./Reference" />
    <!-- second column is the number of parts with that partNum -->
<xsl:value-of select="count(key('pages', ./Reference))" />
 
</xsl:template>


<xsl:template match="/">

<xsl:for-each select="/Call/ObjectSurveys/ObjectSurvey/Results/Result">
<xsl:for-each select=".[not(Reference = preceding::Result/Reference)]/Reference">
<xsl:value-of select="position()"/>
<xsl:value-of select="../Text"/>
<![CDATA[
]]>
</xsl:for-each>
</xsl:for-each>
</xsl:template>      



thanks... I am trying... :)

TBH.
This doesnt answer my question, but my question was so poorly defined that it is a miracle you got me anywhere close. which incidentally I now am.

Excellent work G.

good luck with the 'Sage' level - have 500 points

welcome,
if you need help to close the final gaps, don't hesitate to ask
cheers

ok then, I asked but Im offering the points:

https://www.experts-exchange.com/questions/21881296/to-key-or-path-need-some-RESULTS-nodes-xsl-for-each-select-Call-ObjectSurveys-ObjectSurvey-Survey-contains-Description-'-'.html