I have data that looks like this:


<Stats>


  ...


  <Hits yyyy="2004" mm="02" dd="19" hh="00">


    <FL>76</FL>


    <HP>24</HP>


    <EVNF>27</EVNF>


    <EVS>53</EVS>


  </Hits>


  <Hits yyyy="2004" mm="02" dd="19" hh="01">


    <FL>75</FL>


    <HP>8</HP>


    <CSINF>21</CSINF>


    <CSIS>44</CSIS>


    <EVNF>7</EVNF>


    <EVS>21</EVS>


  </Hits>


  ...


</Stats>


The <Hits /> node covers several months worth of data and will continue to grow. I need to be able to select a range of data within the document so I do the following:


XPathDocument doc = new XPathDocument(Server.MapPath("WebStats.xml"));


XPathNavigator nav = doc.CreateNavigator();


XPathNodeIterator hits = nav.Select("//Hits[concat(@yyyy, @mm, @dd, @hh) >= '" + fromTimestamp + "' and concat(@yyyy, @mm, @dd, @hh) <= '" + thruTimestamp + "']");


Where fromTimestamp and thruTimestamp are values passed in from an aspx page. This works great to get the appropriate <Hits /> for the range that I need. However, I would like to be able to then perform additional "subqueries" on those results. For instance, I want to be able to get the totals (sum) of the various nodes under <Hits />. I know I could switch to using the Evaluate() method and have something along the lines of:


int aaHits = nav.Evaluate("sum(//Hits[concat(@yyyy, @mm, @dd, @hh) >= '" + fromTimestamp + "' and concat(@yyyy, @mm, @dd, @hh) <= '" + thruTimestamp + "']/AA)");


But to do that seems somewhat inefficient having to query the entire document every time for 6 - 8 different sub nodes. In addition, I place the hits object in cache because it is used from a handler that creates chart images based on the data. So ideally I'd like to select the nodes of the date range I need and avoid having to duplicate the logic for filtering in yet another page. I wouldn't mind duplicating the sub-query on another page since it should be simpler.


Thanks in advance for any assistance.


Matt
Matt Smith


Chief Swoof at swoofware.com

Yeah, bad idea. 


If nav is already positioned on appropriate Hits element, why do you need to search it again? Use relative XPath expressions (those don't start with /) - they are evaluated relatively to current node:


XPathNodeIterator hits = nav.Select("//Hits[concat(@yyyy, @mm, @dd, @hh) >= '" + fromTimestamp + "' and concat(@yyyy, @mm, @dd, @hh) <= '" + thruTimestamp + "']"); 


hits.MoveNext();


int aaHits = hits.Current.Evaluate("sum(AA)");

Oleg [XML MVP, MCPD]
XmlLab.Net  |  http://blog.tkachenko.com

double post
Oleg [XML MVP, MCPD]
XmlLab.Net  |  http://blog.tkachenko.com