It looks like it didn't go in yet, you can check what has gone into a
SNAPSHOT by looking at the build server:

https://builds.apache.org/job/Jena__Development_Deploy/changes

I don't see JENA-429 there yet so no this isn't available in a SNAPSHOT yet

I have scheduled a new build but it will be a few hours before the Apache
build servers have a free slot!

Rob


On 4/2/13 4:43 PM, "Diogo FC Patrao" <[email protected]> wrote:

>Hello Rob
>
>I replaced the 2.10.0 library by apache-jena-2.10.1-20130401, however BIND
>aren't getting into my query still.
>
>Is your fix in this version yet?
>
>Thanks
>
>
>--
>diogo patrão
>
>
>
>
>On Tue, Apr 2, 2013 at 1:16 PM, Rob Vesse <[email protected]> wrote:
>
>> The syntax error was correct even in the current stable version
>>
>>
>> The problem was that the code for converting from an algebra back into a
>> query usually puts top level BIND as project expressions which in most
>> cases is semantically equivalent to the original query.  However when
>> there is a GROUP BY present it needs to recreate the BIND as an actual
>> BIND or you get the variable scope error.
>>
>> In algebra terms a project expression and a BIND at the top level of the
>> query pattern are semantically equivalent bar the above proviso about
>> interaction with GROUP BY
>>
>> Rob
>>
>> On 4/1/13 7:47 PM, "Diogo FC Patrao" <[email protected]> wrote:
>>
>> >Hello Rob
>> >
>> >Should now be fixed, build from latest Trunk or wait 24 hrs or so for
>>the
>> >> snapshot to get build and deployed by the Apache build server
>> >>
>> >
>> >* THANKS a million * ! I'm running a couple of experiments for my PhD
>> >thesis, and was getting the excuses done, for that bug messed up part
>>of
>> >my
>> >data. Now I can get everything right. Thanks, thanks.
>> >
>> >
>> >> What version of ARQ are you using to do the OpAsQuery? I assume
>> >> 2.10.1-SNAPSHOT since turning this type of query from its algebra
>>back
>> >> into a query only very recently became possible.  Previously the BIND
>> >> would have been lost completely when converting back from an algebra
>>to
>> >>a
>> >> query.
>> >>
>> >
>> >The jar file says "2.10.0", I just downloaded the latest stable
>>version.
>> >
>> >
>> >> Also is there a reason you are using ARQ 2.8.8 in your system?
>>That's
>> >> very out of date (about 2 years so) now.
>> >>
>> >
>> >well, I'm running this experiment for one year and two months now, and
>> >just
>> >copied the binaries and stuff that were working. TLDR: lazy.
>> >
>> >anyway, 2.10.0 ARQ yielded the same error.
>> >
>> >However, should defining an alias on the query head allow referring to
>>it
>> >elsewhere? Are alias and BINDs really equivalent? MySQL have this
>> >behaviour, but not Oracle.
>> >
>> >Thanks (again).
>> >
>> >dfcp
>> >
>> >
>> >
>> >
>> >> Andy - The report states that the query is passed off to a system
>> >>running
>> >> 2.8.8, even with this fix did ARQ support BIND syntax that far back?
>> >>I.e.
>> >> will this fix just move the error elsewhere?
>> >>
>> >> Rob
>> >>
>> >>
>> >>
>> >> On 4/1/13 12:42 PM, "Andy Seaborne" <[email protected]> wrote:
>> >>
>> >> >Hi there,
>> >> >
>> >> >Thanks for the report. I've captured this as:
>> >> >
>> >> >https://issues.apache.org/jira/browse/JENA-429
>> >> >
>> >> >       Andy
>> >> >
>> >> >On 31/03/13 21:11, Diogo FC Patrao wrote:
>> >> >> Hello
>> >> >>
>> >> >> I was performing a simple test with OpAsQuery:
>> >> >>
>> >> >> Here is the original query:
>> >> >>
>> >> >> PREFIX : <http://www.cipe.accamargo.org.br/ontologias/h2tc.owl#>
>> >> >> PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
>> >> >> PREFIX mylib: <java:dateadd.lib.pkgfor.arq.>
>> >> >>
>> >> >>      SELECT ?yearmonth ( count(?document) as ?total )
>> >> >>      WHERE
>> >> >>      {
>> >> >>          ?document a :Document;
>> >> >>                   :documentDateOfCreation ?date ;
>> >> >>                   :documentType "exam results" .
>> >> >>          BIND( mylib:DateFormat( xsd:string(?date), "yyyy-MM" ) as
>> >> >> ?yearmonth )
>> >> >>      } group by ?yearmonth
>> >> >>
>> >> >> Then I ran the code below, that compile the algebra then render
>>the
>> >> >>query
>> >> >> again.
>> >> >>
>> >> >> Query query = QueryFactory.create( that_query_above,
>> >> >> Syntax.syntaxSPARQL_11);
>> >> >>
>> >> >> Op op = Algebra.compile(query);
>> >> >>
>> >> >> System.out.println(OpAsQuery.asQuery(op));
>> >> >>
>> >> >> The result is below:
>> >> >>
>> >> >> SELECT  (<java:dateadd.lib.pkgfor.arq.DateFormat>(<
>> >> >> http://www.w3.org/2001/XMLSchema#string>(?date), "yyyy-MM") AS
>> >> >>?yearmonth)
>> >> >> (count(?document) AS ?total)
>> >> >>
>> >> >> WHERE
>> >> >>
>> >> >>    { ?document <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <
>> >> >> http://www.cipe.accamargo.org.br/ontologias/h2tc.owl#Document> .
>> >> >>
>> >> >>      ?document <
>> >> >>
>> >> >>
>> >>
>> >>
>> 
>>http://www.cipe.accamargo.org.br/ontologias/h2tc.owl#documentDateOfCreati
>> >> >>on>
>> >> >> ?date .
>> >> >>
>> >> >>      ?document <
>> >> >> http://www.cipe.accamargo.org.br/ontologias/h2tc.owl#documentType>
>> >> "exam
>> >> >> results"
>> >> >>
>> >> >>    }
>> >> >>
>> >> >> GROUP BY ?yearmonth
>> >> >>
>> >> >> However, this is not a valid query; notice that the BIND has
>>gone, it
>> >> >> somehow turned to a alias on query's head.
>> >> >>
>> >> >> ARQ 2.8.8. upon receiving this query, yields the error:
>> >> >>
>> >> >> Variable used when already in-scope: ?yearmonth in
>> >> >> (<java:dateadd.lib.pkgfor.arq.DateFormat>(xsd:string(?date),
>> >>"yyyy-MM")
>> >> >>AS
>> >> >> ?yearmonth)
>> >> >>
>> >> >> Is that a bug on ARQ or OpAsQuery? Or anything else?
>> >> >>
>> >> >> Thanks!
>> >> >>
>> >> >> --
>> >> >> diogo patrão
>> >> >>
>> >> >
>> >>
>> >>
>>
>>

Reply via email to