Peter Kostelnik, PhD.
Mon, 17 May 2010 06:35:13 -0700
hi, ivan .. all right, thanks a lot for a reasonable answer .. strange, but specified .. we'll adapt to this fact ..
thanks once more, cheers,
Peter K.
> Peter,
>
> After a brief discussion of the issue you reported we found out that the
> SPARQL's specification has a special treatment of the empty group pattern:
>
> ---qte--------------------------------------------------------------------------
> 5.2.1 Empty Group Pattern
>
> The group pattern:
>
> { }
> matches any graph (including the empty graph) with one solution that does
> not
> bind any variables. For example:
>
> SELECT ?x
> WHERE {}
> matches with one solution in which variable x is not bound.
> ---eqt--------------------------------------------------------------------------
>
> So in this very case OWLIM correctly follows the spec and this is not a
> bug.
>
> I also find this behavior strange. I believe it originated by SPARQL
> authors'
> desire to handle group patterns like:
>
> {
> ?x foaf:name ?name .
> {}
> ?x foaf:mbox ?mbox .
> }
>
> In such occasions the {} should not affect the whole group (it should be
> ignored). If it did return 0 matches the whole group being an intersection
> of
> the pattern resultsets would return 0 matches.
>
> Anyway..
>
>
> Cheers,
> Ivan
>
>
> On Monday 17 May 2010 15:18:17 Ivan Peikov wrote:
>> Hi Peter,
>>
>> This is indeed a bug. We'll file and address it as soon as possible.
>> Thanks for reporting!
>>
>>
>> Cheers,
>> Ivan
>>
>> On Monday 17 May 2010 14:58:18 Peter Kostelnik, PhD. wrote:
>> > hi there ..
>> >
>> > we've hitted the strange behaviour of BigOWLIM answering the empty
>> query
>> > .. in our case, the query is generated automatically and for this
>> case:
>> >
>> > SELECT ?anything WHERE {}
>> >
>> > the one empty result is returned .. and ?anything is bound to the
>> empty
>> > string ..
>> >
>> > did we find the bug or we are missing something?
>> >
>> > thanks in advance, cheers,
>> > Peter K.
>> >
>> > _______________________________________________
>> > OWLIM-discussion mailing list
>> > OWLIM-discussion@ontotext.com
>> > http://ontotext.com/mailman/listinfo/owlim-discussion
>>
>> _______________________________________________
>> OWLIM-discussion mailing list
>> OWLIM-discussion@ontotext.com
>> http://ontotext.com/mailman/listinfo/owlim-discussion
>
>
> _______________________________________________
> OWLIM-discussion mailing list
> OWLIM-discussion@ontotext.com
> http://ontotext.com/mailman/listinfo/owlim-discussion
>
_______________________________________________
OWLIM-discussion mailing list
OWLIM-discussion@ontotext.com
http://ontotext.com/mailman/listinfo/owlim-discussion