Hi Stephane, Several people have raised the same question. The behavior comes from "stepping up" to a containing element in cases where an element contains an exact match (works well for mixed content, but provides no benefit in this particular case). I agree that the inconsistency is confusing. I'll file a bug to take a look and see if the behavior can be improved for this case.
--Colleen ________________________________________ From: [email protected] [[email protected]] On Behalf Of Stephane Toussaint [[email protected]] Sent: Thursday, July 05, 2012 3:46 AM To: MarkLogic Developer Discussion Subject: [MarkLogic Dev General] search:search matching path coherence Hi, I just find something weird with the matching path returned by the search API. Say I have this xml in my repository <?xml version="1.0" encoding="UTF-8"?> <ARTICLE> <AUTHOR> <NAME>Toussaint</NAME> <FIRSTNAME>Stephane</FIRSTNAME> </AUTHOR> </ARTICLE> If I request this document based on a firstname constraints the document is well returned, but the matching path lake of precision : import module namespace search="http://marklogic.com/appservices/search"; at "/MarkLogic/appservices/search/search.xqy"; let $options := <search:options xmlns:search="http://marklogic.com/appservices/search";> <search:constraint name="firstname"> <search:word> <search:element ns="" name="FIRSTNAME"/> </search:word> </search:constraint> </search:options> return search:search("ste*", $options) <search:response snippet-format="snippet" total="1" start="1" page-length="10" xmlns="" xmlns:search="http://marklogic.com/appservices/search";> <search:result index="1" uri="/content/article/test.xml" path="fn:doc("/content/article/test.xml")" score="49152" confidence="0.371165" fitness="0.677309"> <search:snippet> <search:match path="fn:doc("/content/article/test.xml")/ARTICLE/AUTHOR"> <search:highlight>Stephane</search:highlight> </search:match> </search:snippet> </search:result> [...] </search:response> I expect the path to return fn:doc("/content/article/test.xml")/ARTICLE/AUTHOR/FIRSTNAME but in fact it stops on step upper. But there is inconsistency here because if my firstname was a composed one (2 words), then the returned base is just not the same : <?xml version="1.0" encoding="UTF-8"?> <ARTICLE> <AUTHOR> <NAME>Toussaint</NAME> <FIRSTNAME>Jean Pierre</FIRSTNAME> </AUTHOR> </ARTICLE> <search:response snippet-format="snippet" total="3" start="1" page-length="10" xmlns="" xmlns:search="http://marklogic.com/appservices/search";> <search:result index="1" uri="/content/article/test.xml" path="fn:doc("/content/article/test.xml")" score="32768" confidence="0.303053" fitness="0.680303"> <search:snippet> <search:match path="fn:doc("/content/article/test.xml")/ARTICLE/AUTHOR/FIRSTNAME"><search:highlight>Jean</search:highlight> Pierre</search:match> </search:snippet> </search:result> [...] </search:response> Is it a bug or can someone explain me why search result print different information base on word count from element ? By the way there is in fact no need of constraints to reproduce this exemple a simple full text search provide the very same result. I'd like to provide a filter step based on the path information, and this issue did't allow me to do this. Best Regards Stephane _______________________________________________ General mailing list [email protected] http://developer.marklogic.com/mailman/listinfo/general
