Re: error out of global stack prolog

On 01/09/2014 10:10 AM, Wouter Beek wrote:
> Hi Batselem, Jan,
> 
> It's not straightforward how one would calculate the number of distinct
> triples accros multiple graphs, e.g.:
> ~~~
> ?- rdf_assert(rdf:a, rdf:b, rdf:c, g1).
> ?- rdf_assert(rdf:a, rdf:b, rdf:c, g2).
> ?- rdf_statistics(triples(T)).
> T = 2.
> ~~~
> Or is there another way?

If you want a precise number,

   ?- aggregate_all(count, rdf(_,_,_), Triples).

is the only way.  The statistics give you only an approximation.
There is an internal notion of `possible duplicates' that would
help here, but after removing g2, the triple in g1 is still marked
as a `possible duplicate'.  Note that transaction and snapshot
isolation make the global notion of the number of triples rather
complicated.   The `triples' statistics contains all triples in all
existing `worlds'.  In a particular snapshot, you may see fewer.

But, I don't think that the precise number of triples is a very
interesting property of an RDF graph.

	Cheers --- Jan

> 
> ---
> Cheers!,
> Wouter Beek.
> 
> E-mail: [email protected] <mailto:[email protected]>
> WWW: www.wouterbeek.com <http://www.wouterbeek.com>
> Tel.: 0647674624
> 
> 
> On Thu, Jan 9, 2014 at 9:56 AM, Wielemaker, J. <[email protected]
> <mailto:[email protected]>> wrote:
> 
>     On 01/09/2014 01:59 AM, atom meta wrote:
>     > Hi, I am currently learning Prolog. when I apply a query such as
>     >  findall(A, rdf(A,_,_), Y), length(Y, L), prolog throws out of global
>     > stack, error. How could I fix this problem?
> 
>     You are creating a list with all subjects.  If you have a really big
>     RDF store, that costs a lot of memory.  There are some options:
> 
>       - Raise the stack limit using set_prolog_stack/2.
>       - Use aggregate_all(count, rdf(_,_,_), Triples).
>       - Use rdf_statistics(triples(Triples)).
> 
>     For really counting the triples, the latter is the best (it just
>     looks up the number that is maintained in the store).  In similar
>     cases you may need to do one of the first two.
> 
>             Success --- Jan
> 
>     >
>     > Thank you,
>     > Batselem Jagvaral
>     >
>     >
>     >
>     >
>     > 2 ?- findall(A, rdf(A,_,_), Y), length(Y, L).ERROR: Out of global stack
>     >    Call: (13) rdf_db:rdf_update_duplicates_thread ? No previous search
>     >    Call: (13) rdf_db:rdf_update_duplicates_thread ? No previous search
>     >    Call: (13) rdf_db:rdf_update_duplicates_thread ? creep
>     > ^  Call: (14) thread_create(rdf_update_duplicates, _G75, [detached(true),
>     > alias('__rdf_duplicate_detecter')]) ? creep
>     > ^  Exit: (14) thread_create(rdf_update_duplicates,
>     > '__rdf_duplicate_detecter', [detached(true),
>     > alias('__rdf_duplicate_detecter')]) ?
>     >    Call: (1) rdf_db:rdf_update_duplicates ? creep
>     > |:
>     >    Exit: (13) rdf_db:rdf_update_duplicates_thread ?
>     > -------------- next part --------------
>     > HTML attachment scrubbed and removed
>     > _______________________________________________
>     > SWI-Prolog mailing list
>     > [email protected]
>     <mailto:[email protected]>
>     > https://lists.iai.uni-bonn.de/mailman/listinfo.cgi/swi-prolog
>     >
>     _______________________________________________
>     SWI-Prolog mailing list
>     [email protected]
>     <mailto:[email protected]>
>     https://lists.iai.uni-bonn.de/mailman/listinfo.cgi/swi-prolog
> 
>