Hi, Shruti: MarkLogic calculates facets over indexes. To calculate facets over the filtered result set, the query would have to read every document in the result set on every query, which would scale poorly.
The best approach is to tune the indexes so an unfiltered search closely approximates the desired result set and, if necessary, use filtering to weed out a few exception cases. If the result set is very small, you could extract the XML elements or JSON properties for the facets and do the facet counts in your own code. For most adopters, however, the result set is much too large for that to be a possibility. Erik Hennum ________________________________ From: general-boun...@developer.marklogic.com [general-boun...@developer.marklogic.com] on behalf of shruti kapoor [shrutikapoor....@gmail.com] Sent: Tuesday, January 13, 2015 10:04 PM To: MarkLogic Developer Discussion Subject: Re: [MarkLogic Dev General] Exact match in search:search implementation Hi Erik, I am calculating the total number of results by counting the search:results. For the search facets - the count attribute for each search:facet-value is used, which is not in sync with the search-results count. What could I do create facets of only the filtered search results. Thanks Shruti Kapoor On Tue, Jan 13, 2015 at 8:44 PM, Erik Hennum <erik.hen...@marklogic.com<mailto:erik.hen...@marklogic.com>> wrote: Hi, Shruti: In general, the total is an estimate based on the indexes which will be exactly accurate only if the indexes are turned so the search can run unfiltered. [1] To generate an accurate count for a filtered search, the Search API would have to retrieve and count every document, which would not scale. Can you expand on how you're using the total? You may be able to solve the problem in other ways. For instance, you can terminate paging on the first page with fewer than page-length results. Erik Hennum [1] The exception is that, starting in 8.0, the Search API calculates the total on the last page based on the number of documents paged. ________________________________ From: general-boun...@developer.marklogic.com<mailto:general-boun...@developer.marklogic.com> [general-boun...@developer.marklogic.com<mailto:general-boun...@developer.marklogic.com>] on behalf of shruti kapoor [shrutikapoor....@gmail.com<mailto:shrutikapoor....@gmail.com>] Sent: Tuesday, January 13, 2015 2:00 AM To: MarkLogic Developer Discussion Subject: Re: [MarkLogic Dev General] Exact match in search:search implementation Hi Thanks Peter & David. However I have one more problem. When I apply <search-option>filtered</search-option>, the total search result count and the facet count do not match. How can I return the facet count on the basis of the filtered search results? Thanks Shruti Kapoor On Tue, Jan 6, 2015 at 3:48 PM, David Ennis <david.en...@hinttech.com<mailto:david.en...@hinttech.com>> wrote: HI. Peter, you are correct in what was missing in the search. However, you are right in being confused about the search:qtext..... brace yourself.. Another query console gotcha... Using 'auto' formats it as html and removes the extra whitespace.. When viewed in raw format, your search:report reflects the double space and verifies your solution: Your example expanded to show search parse as well yields the following (expected) whitespace when viewed in raw format: xquery version "1.0-ml"; declare namespace host = "http://marklogic.com/xdmp/status/host";; import module namespace search = "http://marklogic.com/appservices/search"; at "/MarkLogic/appservices/search/search.xqy"; let $options := <options xmlns="http://marklogic.com/appservices/search";> <term> <term-option>case-sensitive</term-option> <term-option>diacritic-sensitive</term-option> <term-option>punctuation-sensitive</term-option> <term-option>whitespace-sensitive</term-option> <term-option>unstemmed</term-option> <term-option>unwildcarded</term-option> </term> <debug>true</debug> <search-option>filtered</search-option> </options> return ( search:search('"Hello world"',$options), search:parse('"Hello world"',$options) ) <search:response snippet-format="snippet" total="0" start="1" page-length="10" xmlns:search="http://marklogic.com/appservices/search";> <search:qtext>"Hello world"</search:qtext> <search:report id="SEARCH-FLWOR">(cts:search(fn:collection(), cts:word-query("Hello world", ("case-sensitive","diacritic-sensitive","punctuation-sensitive","whitespace-sensitive","unstemmed","unwildcarded","lang=en"), 1), ("filtered",cts:score-order("descending")), 1))[1 to 10]</search:report> <search:metrics> <search:query-resolution-time>PT0.00387S</search:query-resolution-time> <search:facet-resolution-time>PT0.00005S</search:facet-resolution-time> <search:snippet-resolution-time>PT0S</search:snippet-resolution-time> <search:total-time>PT0.018433S</search:total-time> </search:metrics> </search:response> <cts:word-query xmlns:cts="http://marklogic.com/cts";> <cts:text xml:lang="en">Hello world</cts:text> <cts:option>case-sensitive</cts:option> <cts:option>diacritic-sensitive</cts:option> <cts:option>punctuation-sensitive</cts:option> <cts:option>whitespace-sensitive</cts:option> <cts:option>unstemmed</cts:option> <cts:option>unwildcarded</cts:option> </cts:word-query> Kind Regards, David Ennis David Ennis Content Engineer [HintTech] <http://www.hinttech.com/> Mastering the value of content creative | technology | content Delftechpark 37i 2628 XJ Delft The Netherlands T: +31 88 268 25 00 M: +31 63 091 72 80 [http://www.hinttech.com]<http://www.hinttech.com> [X] <https://twitter.com/HintTech> [X] <http://www.facebook.com/HintTech> [X] <http://www.linkedin.com/company/HintTech> On 6 January 2015 at 10:06, Peter Kester <peter.kes...@marklogic.com<mailto:peter.kes...@marklogic.com>> wrote: Hi Shruti, MarkLogic does what it is supposed to do. See below sample xquery version "1.0-ml"; xdmp:document-insert("/foo.xml", <foo>Hello world</foo>); xdmp:document-insert("/foo2.xml", <foo>Hello world</foo>); xquery version "1.0-ml"; declare namespace host = "http://marklogic.com/xdmp/status/host";; import module namespace search = "http://marklogic.com/appservices/search"; at "/MarkLogic/appservices/search/search.xqy"; let $options := <options xmlns="http://marklogic.com/appservices/search";> <term> <term-option>case-sensitive</term-option> <term-option>diacritic-sensitive</term-option> <term-option>punctuation-sensitive</term-option> <term-option>whitespace-sensitive</term-option> <term-option>unstemmed</term-option> <term-option>unwildcarded</term-option> </term> <debug>true</debug> <search-option>filtered</search-option> </options> return search:search('"Hello world"',$options) This will give you: <search:response snippet-format="snippet" total="1" start="1" page-length="10" xmlns:xs="http://www.w3.org/2001/XMLSchema"; xmlns=""xmlns:search="http://marklogic.com/appservices/search";> <search:result index="1" uri="/foo.xml" path="fn:doc("/foo.xml")" score="47104" confidence="0.5532626" fitness="0.6769772"> <search:snippet> <search:match path="fn:doc("/foo.xml")/foo"> <search:highlight> Hello world </search:highlight> </search:match> </search:snippet> </search:result> <search:qtext> "Hello world" </search:qtext> <search:report id="SEARCH-FLWOR"> (cts:search(fn:collection(), cts:word-query("Hello world", ("case-sensitive","diacritic-sensitive","punctuation-sensitive","whitespace-sensitive","unstemmed","unwildcarded","lang=en"), 1), ("filtered"), 1))[1 to 10] </search:report> <search:metrics> <search:query-resolution-time> PT0.004736S </search:query-resolution-time> <search:facet-resolution-time> PT0.000051S </search:facet-resolution-time> <search:snippet-resolution-time> PT0.000603S </search:snippet-resolution-time> <search:total-time> PT0.021619S </search:total-time> </search:metrics> </search:response> As you can see it returns the correct document foo.xml and not foo2.xml. The difference is that you also need to specify unfiltered in the options section. I’m not sure why the double spaces are not reflected in the search:qtext or the cts:word-query in the search:report. Hope this helps. Peter Peter Kester Senior Consultant peter.kes...@marklogic.com<mailto:peter.kes...@marklogic.com> +31 611188543<tel:%2B31%20611188543> http://nl.linkedin.com/in/peetkes/ [Description: Description: cid:image001.png@01CCB980.BB82DD90] [Description: Description: MarkLogic Corporation] Graadt van Roggenweg, 328-334, 3531 AH Utrecht http://www.marklogic.com/ New generation databases, you just need to think differently www.nosqlfordummies.com<http://t.co/YKkJ0Wxseo> _______________________________________________ General mailing list General@developer.marklogic.com<mailto:General@developer.marklogic.com> http://developer.marklogic.com/mailman/listinfo/general -- Regards, Shruti Kapoor Software Engineer Innodata India Pvt. Ltd. 7th floor, Stellar IT Park, Sector 62, Noida, Uttar Pradesh 201309, India Cell: (+91) 9990340628 Email: skapo...@innodata.com<http://www.anmsoft.com/> | Web: www.innodata.com<http://www.anmsoft.com/> _______________________________________________ General mailing list General@developer.marklogic.com<mailto:General@developer.marklogic.com> http://developer.marklogic.com/mailman/listinfo/general -- Regards, Shruti Kapoor Software Engineer Innodata India Pvt. Ltd. 7th floor, Stellar IT Park, Sector 62, Noida, Uttar Pradesh 201309, India Cell: (+91) 9990340628 Email: skapo...@innodata.com<http://www.anmsoft.com/> | Web: www.innodata.com<http://www.anmsoft.com/>
_______________________________________________ General mailing list General@developer.marklogic.com http://developer.marklogic.com/mailman/listinfo/general
