W3C home > Mailing lists > Public > public-sparql-dev@w3.org > April to June 2009

Re: query RDF dataset

From: Adrian Walker <adriandwalker@gmail.com>
Date: Fri, 5 Jun 2009 13:24:40 -0400
Message-ID: <1e89d6a40906051024q4363c54cx8b7e7dd1be3eba65@mail.gmail.com>
To: Lee Feigenbaum <lee@thefigtrees.net>, ding@cs.byu.edu, jitao.yang@gmail.com
Cc: public-sparql-dev@w3.org
Jitao, Lee, & All --

You may be interested in the approach to querying RDF in

  www.reengineeringllc.com/demo_agents/RDFQueryLangComparison1.agent

An advantage would appear to be that it is clearer what question is actually
being asked.

BTW, does anyone know please of a reasonably stable SPARQL implementation
that exposes a lower level API than SPARQL itself?

                           Cheers,  -- Adrian

On Fri, Jun 5, 2009 at 12:35 PM, Lee Feigenbaum <lee@thefigtrees.net> wrote:

> Jitao Yang wrote:
>
>> Another problem is:
>> if I add another new book which has no authors to the RDF file like:
>>
>> <?xml version="1.0" encoding="utf-8"?>
>> <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>>         xmlns:dc="http://purl.org/dc/elements/1.1/"
>>     xmlns:vcard="http://www.w3.org/2001/vcard-rdf/3.0#">
>>      <rdf:Description rdf:about="http://example.org/BookJava">
>>        <vcard:fullname>Book Java</vcard:fullname>
>>        <dc:author>
>>            <rdf:Description rdf:about="
>> http://example.org/BookJava/authorA">
>>                 <dc:fullname>Bob Smith</dc:fullname>
>>                 <dc:country>Italy</dc:country>
>>             </rdf:Description>
>>        </dc:author>              <dc:author>
>>            <rdf:Description rdf:about="
>> http://example.org/BookJava/authorB">
>>                 <dc:fullname>Tom Bush</dc:fullname>
>>                 <dc:country>Italy</dc:country>
>>            </rdf:Description>
>>        </dc:author>
>>    </rdf:Description>
>>
>>    <rdf:Description rdf:about="http://example.org/BookCpp">
>>        <vcard:fullname>Book Cpp</vcard:fullname>
>>        <dc:author>
>>            <rdf:Description rdf:about="http://example.org/BookCpp/author2
>> ">
>>                 <dc:fullname>Mike Luck</dc:fullname>
>>                 <dc:country>France</dc:country>
>>            </rdf:Description>
>>        </dc:author>
>>        <dc:author>
>>            <rdf:Description rdf:about="http://example.org/BookCpp/author1
>> ">
>>                 <dc:fullname>Alice Bird</dc:fullname>
>>                 <dc:country>Italy</dc:country>
>>            </rdf:Description>
>>        </dc:author>          <dc:author>
>>            <rdf:Description rdf:about="
>> http://example.org/BookJava/authorB">
>>                 <dc:fullname>kobe Mc</dc:fullname>
>>                 <dc:country>Italy</dc:country>
>>             </rdf:Description>
>>        </dc:author>
>>    </rdf:Description>
>>
>>    <rdf:Description rdf:about="http://example.org/BookC">
>>        <vcard:fullname>Book C</vcard:fullname>
>>    </rdf:Description>
>>
>> </rdf:RDF>
>>
>> I want to query "the books whose authors are all Italian" or "the books
>> which have no authors",
>> so I use SPARQL like:
>>
>>            PREFIX dc:<http://purl.org/dc/elements/1.1/>
>>            PREFIX vcard:<http://www.w3.org/2001/vcard-rdf/3.0#>
>>            SELECT DISTINCT ?bookName
>>            WHERE {
>>                    ?bookURI vcard:fullname ?bookName .
>>                    OPTIONAL{
>>                            ?bookURI dc:author ?author .
>>                            ?author dc:country ?country .
>>                            FILTER(?country = \"Italy\") .
>>                            OPTIONAL{
>>                                    ?bookURI dc:author ?author2 .
>>                                    ?author2 dc:country ?country2 .
>>                                    FILTER(?country2 != \"Italy\")
>>                                    }
>>                            FILTER(bound(?country2))
>>                            }
>>                    FILTER(!bound(?author))
>>                  }
>>
>> I can get the right answers:
>> ---------------
>> | bookName    |
>> ===============
>> | "Book C"    |
>> | "Book Java" |
>> ---------------
>> but, my original idea is FILTER(!bound(?country2)) which means filter the
>> books which have non-Italian authors,
>> however if I use  FILTER(!bound(?country2)), it will return the opposite
>> answers:
>> --------------
>> | bookName   |
>> ==============
>> | "Book C"   |
>> | "Book Cpp" |
>> --------------
>> I can not understand the reasons?
>>
>> Thank you very much if Lee would like to explain to me,
>> or the other persons who have time to help me?
>>
>
> Because of the !bound(?author), you're only going to receive books that
> don't satisfy your outer optional. Because the outer optional requires an
> author, this will include books with no author, as you want.
>
> Books with an author will only satisfy the outer optional if
> bound(?country2) is true, which is true only if the book has a non-Italian
> author. So you end up excluding those as well, as you want.
>
> That said, I think you've made the query overly complicated. Would this
> accomplish what you want?
>
>            PREFIX dc:<http://purl.org/dc/elements/1.1/>
>            PREFIX vcard:<http://www.w3.org/2001/vcard-rdf/3.0#>
>            SELECT DISTINCT ?bookName
>            WHERE {
>                    ?bookURI vcard:fullname ?bookName .
>                    OPTIONAL{
>                            ?bookURI dc:author ?author .
>                            ?author dc:country ?country .
>                            FILTER(?country != "Italy") .
>                    }
>                    FILTER(!bound(?author))
>           }
>
> (This is simply the original query but with the part that required at least
> one Italian author removed.)
>
>
> Lee
>
>  Best wishes,
>> Jitao
>>
>>
>> On Sat, May 30, 2009 at 21:51, Lee Feigenbaum <lee@thefigtrees.net<mailto:
>> lee@thefigtrees.net>> wrote:
>>
>>    Jitao Yang wrote:
>>
>>        Dear all,
>>
>>        if there is a RDF like:
>>
>>        <?xml version="1.0" encoding="utf-8"?>
>>        <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>>                xmlns:dc="http://purl.org/dc/elements/1.1/">
>>           <rdf:Description rdf:about="http://example.org/BookJava">
>>               <dc:fullname>Book Java</dc:fullname>
>>               <dc:author>
>>                   <rdf:Description
>>        rdf:about="http://example.org/BookJava/authorA">
>>                        <dc:fullname>Bob Smith</dc:fullname>
>>                        <dc:country>Italy</dc:country>
>>                   </rdf:Description>
>>               </dc:author>              <dc:author>
>>                   <rdf:Description
>>        rdf:about="http://example.org/BookJava/authorB">
>>                        <dc:fullname>Tom Bush</dc:fullname>
>>                        <dc:country>Italy</dc:country>
>>                   </rdf:Description>
>>               </dc:author>
>>           </rdf:Description>
>>
>>           <rdf:Description rdf:about="http://example.org/BookCpp">
>>               <dc:fullname>Book Cpp</dc:fullname>
>>               <dc:author>
>>                   <rdf:Description
>>        rdf:about="http://example.org/BookCpp/author1">
>>                        <dc:fullname>Alice Bird</dc:fullname>
>>                        <dc:country>Italy</dc:country>
>>                   </rdf:Description>
>>               </dc:author>          <dc:author>
>>                   <rdf:Description
>>        rdf:about="http://example.org/BookCpp/author2">
>>                        <dc:fullname>Mike Luck</dc:fullname>
>>                        <dc:country>France</dc:country>
>>                   </rdf:Description>
>>               </dc:author>
>>           </rdf:Description>
>>        </rdf:RDF>
>>
>>
>>        and I query the RDF by:
>>
>>        PREFIX dc:<http://purl.org/dc/elements/1.1/>
>>        SELECT DISTINCT ?bookName
>>        WHERE {
>>                         ?bookName dc:author ?author .
>>                         {?author dc:country ?country . FILTER(?country
>>        = "Italy")}
>>                     };
>>
>>        the query results are:
>>
>>        ---------------------------------
>>        | bookName                      |
>>        =================================
>>        | <http://example.org/BookCpp>  |
>>        | <http://example.org/BookJava> |
>>        ---------------------------------
>>
>>        can anybody tell me how to query out the "book whose authors'
>>        country are all Italy "?
>>
>
>
>
>
>         which means the query results should be:
>>
>>        ---------------------------------
>>        | bookName                      |
>>        =================================
>>        | <http://example.org/BookJava> |
>>        ---------------------------------
>>
>>
>>    Jitao,
>>
>>    It's not pretty, but you should be able to do it with something like:
>>
>>    PREFIX dc:<http://purl.org/dc/elements/1.1/>
>>    SELECT DISTINCT ?bookName
>>    WHERE {
>>     # first, make sure there's an author from italy
>>     ?bookName dc:author ?author .
>>
>>     ?author dc:country ?country . FILTER(?country = "Italy") .
>>     # next, "try" to find another author not from italy
>>     OPTIONAL {
>>       ?bookName dc:author ?author2 .
>>       ?author2 dc:country ?country2 .
>>       FILTER(?country2 != "Italy")
>>     }
>>     # finally, we only want this book if we didn't find any non-Italy
>>     # authors
>>     FILTER(!bound(?country2))
>>    }
>>
>>    (Warning, untested query.)
>>
>>    hope this helps,
>>    Lee
>>
>>
>>
>>        Thank you very much for your help!
>>
>>        Best wishes,
>>        Jitao
>>
>>
>>
>>
>
Received on Friday, 5 June 2009 17:25:22 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 5 June 2009 17:25:22 GMT