On 5/4/12 9:08 AM, Alex wrote:
Thanks I will try that. Though with a quick view I didn't understand it. Are
there any instance data stored? Which graphs are already stored?

Again I get errors searching for graphs using
http://demo.openlinksw.com/isparql/:

Virtuoso ..... Error SQ200: Query too large, more than 65000 variables in state

There is a link on /sparql .

Example link: http://lod.openlinksw.com/sparql?rdfinf (preferred instance for this exercise since it has the most resources behind it).
Rule Name for RDFS Schema: "http://www.w3.org/2000/01/rdf-schema"; .

Kingsley

I think our dataset and possible the schema on which the rule is based
have a copyright or some other restrictions and I might not be able to
share them.I am going to contact the
person responsible and let you know.

On 4 May 2012 04:21, Kingsley Idehen<[email protected]>  wrote:
On 5/3/12 1:08 PM, Hugh Williams wrote:

Hi Alex,

Are you able to provide some sample data for the inference rule you are
creating and the datasets it is being applied against so we can follow a set
of steps to recreate this problem in-house ?  Or if you can recreate on one
of our public SPARQL endpoints that would be helpful also as in both case we
have a test case to look into and resolve the issue if required ?


Alex,

In our earlier exchanges, I forgot to mention to you the fact that a number
of our public facing sparql endpoints can be used as problem recreation
beds. In your case, assuming you are using the RDFS Schema as your inference
rule, you can simply make use of our preset inference rules. Here's what you
need to do:

1. Goto to one of the endpoints
2. Select an inference rule
3. Use the selected inference rule in a sparql query scoped to an existing
named graph IRI or use the sponger feature (via FROM CLAUSE (option get:soft
"add" ) or define get:soft "add" pragma) .

Possible endpoints:

1. http://demo.openlinksw.com/sparql
2. http://uriburner.com/sparql .

Rule name for RDFS Schema is "RDFs" , and its based on the ontology at:
http://www.w3.org/2000/01/rdf-schema . Thus, you are going to be executing:

Data Creation:

sparql
define get:soft "add"
define input:inference "RDFs"
insert {graph<rdfs_g>  {?s ?p ?o}}
using<g>  where {?s ?p ?o}

and

sparql
define get:soft "add"
define input:inference "RDFs"
insert {graph<rdfs_g>  {<uri>  ?p ?o}}
using<g>  where {<uri>  ?p ?o}

then make your sample queries that demonstrate the problem in question.

Kingsley






Best Regards
Hugh Williams
Professional Services
OpenLink Software, Inc.      //              http://www.openlinksw.com/
10 Burlington Mall Road, Suite 265, Burlington MA 01803
Weblog   -- http://www.openlinksw.com/blogs/
LinkedIn -- http://www.linkedin.com/company/openlink-software/
Twitter  -- http://twitter.com/OpenLink
Google+  -- http://plus.google.com/100570109519069333827/
Facebook -- http://www.facebook.com/OpenLinkSoftware
Universal Data Access, Integration, and Management Technology Providers

On 3 May 2012, at 15:19, Alex wrote:

Hello,
so the referred problem also exists in the 6.1.5 version.
This time I tried the following queries w.r.t Sparql 1.1:

sparql define input:inference<something>  insert {graph<rdfs_g>{<uri>
?p ?o}} using<g>  where{<uri>  ?p ?o}; (1)

sparql define input:inference<something>  insert {graph<rdfs_g>{?s ?p
?o}} using<g>  where{?s ?p ?o}; (2)

Using the following query:
sparql select * from<rdfs_g>  where{<uri>  ?p ?o};

The result after (1) is different than those after (2).

The tests regards the number of inferred triples and a requirement is
backward chaining.
I cannot have a "safe" outcome with the above behavior.

I know that virtuoso is huge and maybe requires a lot of work.
However, I am not so sure if I have enough time anymore.
The deadlines are chasing me and I am afraid I should change triple store.


On 27 April 2012 18:52, Kingsley Idehen<[email protected]>  wrote:

On 4/27/12 10:18 AM, Alex wrote:


Thank you for your support.


Two more things to take into consideration on the previous statistics.


Even though the total number of triples in<rdfs_g>    and<g>    might be 8

and that there is 1 class instance in both graphs this does not mean

that the triples are the same in<rdfs_g>    and<g>,

e.g., the triple (<uri>    rdf:type ?o) is different in<g>    and in

<rdfs_g>    w.r.t ?o (i.e., 2 different class instances, where in

<rdfs_g>    is only one inferred).


Moreover, I have installed 06.01.327 version on ubuntu. Meanwhile I do not

promise that I will upgrade virtuoso (maybe only after an official

response) because of the possible danger of

messing up with the dataset and/or current configuration.



Please upgrade as this helps resolution of this matter.


You can dump your dataset and then reload into the upgraded instance.


You can also make backup and restore.



Kingsley



On 27 April 2012 16:57, Kingsley Idehen<[email protected]>    wrote:


On 4/27/12 9:31 AM, Alex wrote:


Yes rdfs schema.


15 is the total number of triples including both the inferred and the

explicit ones from<g>.



Okay, I am going to hand this over to our support team to see if they can

recreate based on the information in this post. Naturally, you have to

bear

with the support turnaround which is always based on what's on the queue

etc..


They (or I) will get back to you.



Kingsley



On 27 April 2012 15:45, Kingsley Idehen<[email protected]>

  wrote:


On 4/27/12 8:36 AM, Alex wrote:


I can give you some statistics if this is helpful.


Number of triples in<g>: 152329

Number of triples with subject<uri>        in<g>: 8

Number of triples with subject<uri>        in<rdfs_g>        : 8

Number of triples applying rdfs on triples with subject<uri>

  in<g>:

15



I assume you are using rdfs schema for your inference rule? If true, I

also

assume 15 is the inferred triples count?



Kingsley


sparql define input:inference<something>        select count(*) from<g>

where{?s ?p ?o} : 152329 !

sparql define input:inference<something>        select count(*) from<g>

where{<uri>        rdf:type ?o} : 1

sparql define input:inference<something>        select count(*)

from<rdfs_g>

where{<uri>        rdf:type ?o}: 1


sparql clear graph<rdfs_g>

sparql define input:inference<something>        insert into<rdfs_g>

{<uri>?p ?o} where{graph<g>        {<uri>        ?p ?o}}

sparql define input:inference<something>        select count(*) from

<rdfs_g>        where{<uri>        rdf:type ?o}: 6


I hope there will be no typos. Do you need anything else to share with

you? Please let me know.


Thanks.


On 27 April 2012 15:05, Kingsley Idehen<[email protected]>

  wrote:


On 4/27/12 7:35 AM, Alex wrote:


Following the Carina's advice:


shutdown()

change value of parameter "ResultSetMaxRows" in the .ini file to

1000000000000000

save

sudo virtuoso-t -f&

sudo isql-vt


Applying again the rule, I get same the wrong behavior: inference

only

on the<uri>          results to the correct inferred triples, but not

in

the

case of inference on the whole dataset.


I think and I am afraid I cannot "share" the dataset I am testing (

if

you meant that).


I changed my query to selection:


sparql define input:inference<something>          select * from<g>

  where{?s

?p

?o} limit 20


I do not know how the inference is evaluated on the triples and how

the result is returned, but in the first 20 results I do not see any

new inferred triple.



Without sharing a SPARQL URL you can share the results of a SELECT

query

e.g., a count of the records when using a SELECT query.


Kingsley


On 27 April 2012 14:18, Kingsley Idehen<[email protected]>

  wrote:


On 4/27/12 6:57 AM, Carina Haupt wrote:



On 27.04.2012 12:14, Alex wrote:


Hello again,

my scenario is the following: given a graph<g>             with

some

triples

apply a rule set based on a schema and insert the inferred

triples

to

another graphs<rdfs_g>.


So I use the following query to accomplish that:


sparql define input:inference<something>             insert

into<rdfs_g>

{?s ?p

?o} where{graph<g>             {?s ?p ?o}} (1)


I then check for a specific uri in rdfs_g and I see the same

inferred

triples but NOT all of them and the number of triples is the

same.


To further test it I apply the same rule for this specific uri:


sparql define input:inference<something>             insert

into<rdfs_g>

{<uri>

?p ?o} where{graph<g>             {<uri>             ?p ?o}} (2)


Comparing the<rdfs_g>             to<g>             there is big

difference

on

the

number of

triples (i.e. 15 to 8 rows respectively) and I think I see the

total

inferred triples.


What is wrong with that?? I have test it (1) on a smaller dataset

that

<g>             and works fine.


I am begging to lose my mind on this. How could I debug this?


There is a threshold in the ini file which defines how many

triples

are

returned at max. Perhaps this is the source of your problem?!


Ciao Carina



Yes re. INI.


Alternatively, one diagnostic option re. assistance from us would

be

to

change the INSERT into a SELECT and then share the SPARQL protocol

URL.


Kingsley


Thanks.







------------------------------------------------------------------------------

Live Security Virtual Conference

Exclusive live event will cover all the ways today's security and

threat landscape has changed and how IT managers can respond.

Discussions

will include endpoint security, mobile security and the latest in

malware

threats.

http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

_______________________________________________

Virtuoso-users mailing list

[email protected]

https://lists.sourceforge.net/lists/listinfo/virtuoso-users




--


Regards,


Kingsley Idehen

Founder&            CEO

OpenLink Software

Company Web: http://www.openlinksw.com

Personal Weblog: http://www.openlinksw.com/blog/~kidehen

Twitter/Identi.ca handle: @kidehen

Google+ Profile:

https://plus.google.com/112399767740508618350/about

LinkedIn Profile: http://www.linkedin.com/in/kidehen












------------------------------------------------------------------------------

Live Security Virtual Conference

Exclusive live event will cover all the ways today's security and

threat landscape has changed and how IT managers can respond.

Discussions

will include endpoint security, mobile security and the latest in

malware

threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

_______________________________________________

Virtuoso-users mailing list

[email protected]

https://lists.sourceforge.net/lists/listinfo/virtuoso-users


--


Regards,


Kingsley Idehen

Founder&          CEO

OpenLink Software

Company Web: http://www.openlinksw.com

Personal Weblog: http://www.openlinksw.com/blog/~kidehen

Twitter/Identi.ca handle: @kidehen

Google+ Profile: https://plus.google.com/112399767740508618350/about

LinkedIn Profile: http://www.linkedin.com/in/kidehen








------------------------------------------------------------------------------

Live Security Virtual Conference

Exclusive live event will cover all the ways today's security and

threat landscape has changed and how IT managers can respond.

Discussions

will include endpoint security, mobile security and the latest in

malware

threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

_______________________________________________

Virtuoso-users mailing list

[email protected]

https://lists.sourceforge.net/lists/listinfo/virtuoso-users




--


Regards,


Kingsley Idehen

Founder&        CEO

OpenLink Software

Company Web: http://www.openlinksw.com

Personal Weblog: http://www.openlinksw.com/blog/~kidehen

Twitter/Identi.ca handle: @kidehen

Google+ Profile: https://plus.google.com/112399767740508618350/about

LinkedIn Profile: http://www.linkedin.com/in/kidehen










------------------------------------------------------------------------------

Live Security Virtual Conference

Exclusive live event will cover all the ways today's security and

threat landscape has changed and how IT managers can respond.

Discussions

will include endpoint security, mobile security and the latest in

malware

threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

_______________________________________________

Virtuoso-users mailing list

[email protected]

https://lists.sourceforge.net/lists/listinfo/virtuoso-users



------------------------------------------------------------------------------

Live Security Virtual Conference

Exclusive live event will cover all the ways today's security and

threat landscape has changed and how IT managers can respond.

Discussions

will include endpoint security, mobile security and the latest in

malware

threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

_______________________________________________

Virtuoso-users mailing list

[email protected]

https://lists.sourceforge.net/lists/listinfo/virtuoso-users




--


Regards,


Kingsley Idehen

Founder&      CEO

OpenLink Software

Company Web: http://www.openlinksw.com

Personal Weblog: http://www.openlinksw.com/blog/~kidehen

Twitter/Identi.ca handle: @kidehen

Google+ Profile: https://plus.google.com/112399767740508618350/about

LinkedIn Profile: http://www.linkedin.com/in/kidehen









------------------------------------------------------------------------------

Live Security Virtual Conference

Exclusive live event will cover all the ways today's security and

threat landscape has changed and how IT managers can respond. Discussions

will include endpoint security, mobile security and the latest in malware

threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

_______________________________________________

Virtuoso-users mailing list

[email protected]

https://lists.sourceforge.net/lists/listinfo/virtuoso-users



------------------------------------------------------------------------------

Live Security Virtual Conference

Exclusive live event will cover all the ways today's security and

threat landscape has changed and how IT managers can respond. Discussions

will include endpoint security, mobile security and the latest in malware

threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

_______________________________________________

Virtuoso-users mailing list

[email protected]

https://lists.sourceforge.net/lists/listinfo/virtuoso-users




--


Regards,


Kingsley Idehen

Founder&    CEO

OpenLink Software

Company Web: http://www.openlinksw.com

Personal Weblog: http://www.openlinksw.com/blog/~kidehen

Twitter/Identi.ca handle: @kidehen

Google+ Profile: https://plus.google.com/112399767740508618350/about

LinkedIn Profile: http://www.linkedin.com/in/kidehen








------------------------------------------------------------------------------

Live Security Virtual Conference

Exclusive live event will cover all the ways today's security and

threat landscape has changed and how IT managers can respond. Discussions

will include endpoint security, mobile security and the latest in malware

threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

_______________________________________________

Virtuoso-users mailing list

[email protected]

https://lists.sourceforge.net/lists/listinfo/virtuoso-users



------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Virtuoso-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/virtuoso-users




------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/



_______________________________________________
Virtuoso-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/virtuoso-users



--

Regards,

Kingsley Idehen 
Founder&  CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen





------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Virtuoso-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/virtuoso-users

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Virtuoso-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/virtuoso-users


--

Regards,

Kingsley Idehen 
Founder&  CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen






Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to