Joshil, Yes, the @total attribute will return the *estimated* total number of documents (or fragments) that match the query, not the number of matching terms in the document. The @total is a quick estimate: it can only give as much information as can be resolved directly from indexes. In some circumstances, like yours, it won’t match the number of <result> elements that are actually returned, because there isn’t enough information in the index to make the estimate completely accurate.
In your case, the searchable expression is the equivalent of saying: find me <title> elements that contain “wine” (without the searchable expression, you’d be saying “find me documents that contain “wine”). Your document is indexed at the <root> level, not the <title> level. So from the indexes, the sever knows that the document contains *at least one* <title> element that contains “wine”; that’s the best estimate it can make. To be more accurate, the server has to do some filtering to really resolve whether it’s one, or more than one. In this case, it’s 2. So your results are accurate (they’ve been through the filter) but the @total is not (it was just a quick, unfiltered estimate). This is a fairly complicated topic: I recommend taking a look at the section on “Fast Pagination and Unfiltered Search” in this guide: http://developer.marklogic.com/pubs/4.1/books/performance.pdf for some hints on how things work under the covers. --Colleen From: [email protected] [mailto:[email protected]] On Behalf Of Joshil Avikkal Sent: Thursday, August 06, 2009 10:48 AM To: Joshil Avikkal; General Mark Logic Developer Discussion Subject: RE: [MarkLogic Dev General] pagination using new search API I forgot to add this… when I searched for “wine”, the search returned 2 result elements but total shows ‘1’. I have “searchable-expression” option set to “//title”. --Joshil ________________________________ From: Joshil Avikkal Sent: Thursday, August 06, 2009 1:40 PM To: 'General Mark Logic Developer Discussion' Subject: RE: [MarkLogic Dev General] pagination using new search API It appears that the total attribute always returns me the number of documents not the total number of search result. When searched for “wine” in below example, I got total=”1” when I was expecting total=” 2”. Am I missing something? --Joshil ________________________________ From: [email protected] [mailto:[email protected]] On Behalf Of Colleen Whitney Sent: Thursday, August 06, 2009 1:23 PM To: General Mark Logic Developer Discussion Subject: RE: [MarkLogic Dev General] pagination using new search API The total attribute in response gives you the estimated total number of results for the search. The number of results actually contained in the response wrapper is controlled by page-length, which can be passed as an option or as an argument to search:search. (In other words, search:search() paginates results.) --Colleen From: [email protected] [mailto:[email protected]] On Behalf Of Joshil Avikkal Sent: Thursday, August 06, 2009 10:11 AM To: Joshil Avikkal; General Mark Logic Developer Discussion Subject: RE: [MarkLogic Dev General] pagination using new search API Thanks Colleen. This answers most of my questions. Which element or attribute in <response> or its child nodes to point out total number of <result> elements returned by the search? --Joshil ________________________________ From: [email protected] [mailto:[email protected]] On Behalf Of Colleen Whitney Sent: Thursday, August 06, 2009 12:36 PM To: General Mark Logic Developer Discussion Subject: RE: [MarkLogic Dev General] pagination using new search API Hi Joshil, I’m not sure I completely understand what you want to do, but here are a couple of ideas: Each <result> that comes back from search:search represents a document or fragment in your database. So if you have one sample document in your database (and it isn’t fragmented), then search:search(“wine”) will return just one <result> element. Using <max-matches> only controls the number of matching nodes returned as part of the snippet; it won’t change the number of <result> elements that are returned. In order to restrict the search to terms in the <title> element, you need to construct an options node that defines a word constraint on the title element; that will allow you to write a search like: search:search(“title:wine”) and get documents that have “wine” only in the title element. The result element includes a document URI and path, so that you can access the document and use xpath on it to get the value of the <title> tag, which you can then display along with the snippet. There is currently no way to include the value of the <title> element with your snippet (there is an RFE on file for that). I’d like to recommend that you try going through the following tutorial, which gives you a hands-on tour of the Search API, including how to set up constraints and use some of the other basic features. http://developer.marklogic.com/howto/tutorials/2009-07-search-api-walkthrough.xqy I hope this helps! Regards, --Colleen Whitney From: [email protected] [mailto:[email protected]] On Behalf Of Joshil Avikkal Sent: Thursday, August 06, 2009 7:55 AM To: General Mark Logic Developer Discussion Subject: RE: [MarkLogic Dev General] pagination using new search API Hi, I have an xml like follows and my database contains just one document with similar tags. <Root> <title>cook1 <Category count =”1” >wine and beer, wine and water </category <Category count =”2” >this is a sample document </category> </title> <title>cook2 <Category count =”1” >party and cocktail</category> <title> <title>cook3 <Category count =”1” >wine</category> </title> </root> I would like to use “search: search” function and get following the string term “wine”. . 1. Total number of the matches found in number of <title> tags. In this example, although the “wine” appears twice in the <title> “Cook1”, it should be counted as 1 and it should return only two search:result nodes. I set <max-matches> to 1 control this. But I need to find out how many such matches exist in entire document. Is there way to find this out without traversing through the entire search:result ? 2. Along with the search:snippet I always want to display the value of the <title> tag inside with search:result along with search:snippet. How can I do this ? Thank you, --Joshil --------------------------------------------------------------------------------------------- This message, including any attachments, contains confidential information intended for a specific individual and purpose, and is intended for the addressee only. Any unauthorized disclosure, use, dissemination, copying, or distribution of this message or any of its attachments or the information contained in this e-mail, or the taking of any action based on it, is strictly prohibited. If you are not the intended recipient, please notify the sender immediately by return e-mail and delete this message. --------------------------------------------------------------------------------------------- -------------------------------------------------------------------------------------------- This message, including any attachments, contains confidential information intended for a specific individual and purpose, and is intended for the addressee only. Any unauthorized disclosure, use, dissemination, copying, or distribution of this message or any of its attachments or the information contained in this e-mail, or the taking of any action based on it, is strictly prohibited. If you are not the intended recipient, please notify the sender immediately by return e-mail and delete this message. -------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------- This message, including any attachments, contains confidential information intended for a specific individual and purpose, and is intended for the addressee only. Any unauthorized disclosure, use, dissemination, copying, or distribution of this message or any of its attachments or the information contained in this e-mail, or the taking of any action based on it, is strictly prohibited. If you are not the intended recipient, please notify the sender immediately by return e-mail and delete this message. --------------------------------------------------------------------------------------------- --------------------------------------------------------------------------------------------- This message, including any attachments, contains confidential information intended for a specific individual and purpose, and is intended for the addressee only. Any unauthorized disclosure, use, dissemination, copying, or distribution of this message or any of its attachments or the information contained in this e-mail, or the taking of any action based on it, is strictly prohibited. If you are not the intended recipient, please notify the sender immediately by return e-mail and delete this message. ---------------------------------------------------------------------------------------------
_______________________________________________ General mailing list [email protected] http://xqzone.com/mailman/listinfo/general
