Re: Question about RDF collections

[Laurent Rucquoy]

Thank you all for your help !

On 22 September 2016 at 14:20, Nikolaos Beredimas  wrote:

> Just remember that there is no ordering in this.
>
> [Document_A] --contains--> [Paragraph_1]
> [Document_A] --contains--> [Paragraph_2]
> [Document_B] --contains--> [Paragraph_2]
>
> is equivalent to
>
> [Document_A] --contains--> [Paragraph_2]
> [Document_B] --contains--> [Paragraph_2]
> [Document_A] --contains--> [Paragraph_1]
>
>
> On Thu, Sep 22, 2016 at 3:14 PM, Laurent Rucquoy <
> laurent.rucq...@telemis.com> wrote:
>
> > I have a "Document" resource which could contain many "Paragraph"
> > resources.
> > A same "Paragraph" resource could also be contained by different
> "Document"
> > resources.
> >
> > What is the most relevant model to translate such a case in RDF ?
> > I have two solutions:
> >
> > Solution 1 (using RDF collections)
> > [Document_A] --contains--> ( [Paragraph_1], [Paragraph_2] )
> > [Document_B] --contains--> ( [Paragraph_2] )
> >
> > or
> >
> > Solution 2 (defining the same predicate several times on the same
> subject)
> > [Document_A] --contains--> [Paragraph_1]
> > [Document_A] --contains--> [Paragraph_2]
> > [Document_B] --contains--> [Paragraph_2]
> >
> >
> > I think that the Solution 1 requires complex and resource-consuming
> SPARQL
> > update.
> > So to keep it simple, I would choose the Solution 2. But, I don't know if
> > it's safe and if it's a good practice to define the same predicate
> several
> > times on the same subject ?
> >
> > Thank you in advance for your help.
> >
> > Regards,
> > Laurent
> >
>

Re: Question about RDF collections

[Nikolaos Beredimas]

Just remember that there is no ordering in this.

[Document_A] --contains--> [Paragraph_1]
[Document_A] --contains--> [Paragraph_2]
[Document_B] --contains--> [Paragraph_2]

is equivalent to

[Document_A] --contains--> [Paragraph_2]
[Document_B] --contains--> [Paragraph_2]
[Document_A] --contains--> [Paragraph_1]


On Thu, Sep 22, 2016 at 3:14 PM, Laurent Rucquoy <
laurent.rucq...@telemis.com> wrote:

> I have a "Document" resource which could contain many "Paragraph"
> resources.
> A same "Paragraph" resource could also be contained by different "Document"
> resources.
>
> What is the most relevant model to translate such a case in RDF ?
> I have two solutions:
>
> Solution 1 (using RDF collections)
> [Document_A] --contains--> ( [Paragraph_1], [Paragraph_2] )
> [Document_B] --contains--> ( [Paragraph_2] )
>
> or
>
> Solution 2 (defining the same predicate several times on the same subject)
> [Document_A] --contains--> [Paragraph_1]
> [Document_A] --contains--> [Paragraph_2]
> [Document_B] --contains--> [Paragraph_2]
>
>
> I think that the Solution 1 requires complex and resource-consuming SPARQL
> update.
> So to keep it simple, I would choose the Solution 2. But, I don't know if
> it's safe and if it's a good practice to define the same predicate several
> times on the same subject ?
>
> Thank you in advance for your help.
>
> Regards,
> Laurent
>

Re: Question about RDF collections

[Martynas Jusevičius]

It is not only safe, but completely normal and used much, much more
frequently than solution 1. Remember, RDF is a graph, so you can
connect resource nodes however you like.

Collection could be useful here if you want to maintain the ordering
of Paragraphs in a Document.

On Thu, Sep 22, 2016 at 2:14 PM, Laurent Rucquoy
 wrote:
> I have a "Document" resource which could contain many "Paragraph" resources.
> A same "Paragraph" resource could also be contained by different "Document"
> resources.
>
> What is the most relevant model to translate such a case in RDF ?
> I have two solutions:
>
> Solution 1 (using RDF collections)
> [Document_A] --contains--> ( [Paragraph_1], [Paragraph_2] )
> [Document_B] --contains--> ( [Paragraph_2] )
>
> or
>
> Solution 2 (defining the same predicate several times on the same subject)
> [Document_A] --contains--> [Paragraph_1]
> [Document_A] --contains--> [Paragraph_2]
> [Document_B] --contains--> [Paragraph_2]
>
>
> I think that the Solution 1 requires complex and resource-consuming SPARQL
> update.
> So to keep it simple, I would choose the Solution 2. But, I don't know if
> it's safe and if it's a good practice to define the same predicate several
> times on the same subject ?
>
> Thank you in advance for your help.
>
> Regards,
> Laurent

Re: Question about RDF collections

[Joshua TAYLOR]

On Thu, Sep 22, 2016 at 8:14 AM, Laurent Rucquoy
 wrote:
> But, I don't know if
> it's safe and if it's a good practice to define the same predicate several
> times on the same subject ?

It's absolutely fine.  That's how, e.g., DBpedia stores abstracts in
different languages:

article dbpedia-owl:abstract "abstract in english"@en .
article dbpedia-owl:abstract "abstract in french"@fr .
article dbpedia-owl:abstract "abstract in italian"@it .



-- 
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/

Question about RDF collections

[Laurent Rucquoy]

I have a "Document" resource which could contain many "Paragraph" resources.
A same "Paragraph" resource could also be contained by different "Document"
resources.

What is the most relevant model to translate such a case in RDF ?
I have two solutions:

Solution 1 (using RDF collections)
[Document_A] --contains--> ( [Paragraph_1], [Paragraph_2] )
[Document_B] --contains--> ( [Paragraph_2] )

or

Solution 2 (defining the same predicate several times on the same subject)
[Document_A] --contains--> [Paragraph_1]
[Document_A] --contains--> [Paragraph_2]
[Document_B] --contains--> [Paragraph_2]


I think that the Solution 1 requires complex and resource-consuming SPARQL
update.
So to keep it simple, I would choose the Solution 2. But, I don't know if
it's safe and if it's a good practice to define the same predicate several
times on the same subject ?

Thank you in advance for your help.

Regards,
Laurent