Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[protege-owl] MappingMaster: specify ID

Roland Cornelissen roland at metamatter.nl
Tue Dec 14 12:12:00 PST 2010


Exactly what I have here at hand; non unique labels.
And composing an unique ID doesn't do the trick for me. The result is
always 1 individual differentiated by it's properties, while I wish it
were multiple individuals each with their own properties.
My guess is this can only be done by defining ID's in my source data and
not defining ID's when importing the individuals(?)


On 12/14/2010 08:56 PM, Martin O'Connor wrote:
> 
> Hopefully the composition functionality should suffice.
> 
> Another reason that generating the ID of an individual from the
> associated class would be problematic is that individuals can also be
> declared using class expressions and not just named classes, e.g.,
> 
> Individual: @A3(rdfs:label) Type: p1 some ClassA and p2 value "323"
> 
> Dealing with ontologies that contain non unique labels is always going
> to be problematic. I will try to come up with a solution because this
> situation is relatively common.
> 
> Martin
> 
> On 12/14/2010 11:28 AM, Roland Cornelissen wrote:
>> I see. I hoped that maybe the feature that Protege already has could be
>> used; starting the ID of an added individual with the classname it is
>> created in. So that would just count for individuals.
>> But with the feature of MM to compose strings in IDs I think I have
>> enough for now.
>>
>> Thanks!
>> Roland
>>
>>> Supporting this feature would seriously complicate the language because
>>> it essentially would make reference processing context sensitive. For
>>> example, the reference @A3 in the following two expressions
>>>
>>> Individual: @A3(rdfs:label) Type: ClassA
>>>
>>> Individual: @A3(rdfs:label) Type: ClassB
>>>
>>> would have to be processed differently based on its surrounding
>>> expression, so @A3(rdfs:label) would mean different things based on
>>> where it was used.
>>>
>>> By design, reference processing is not context sensitive and MM ensures
>>> that the same reference used in different expressions renders the same
>>> entity.
>>>
>>> Perhaps there is an alternative approach of doing what you are trying to
>>> do. I am open to suggestions. However, as a general point, rdf:IDs
>>> should not be meaningful - or at least the IDs should not carry any
>>> semantics.
>>>
>>> Martin
>>>
>>> On 12/13/2010 11:25 AM, Roland Cornelissen wrote:
>>>> I have formulated my question wrong, sorry.
>>>> The situation is I have existing classes with different names. I try to
>>>> map individuals to these classes, where individuals can have the same
>>>> label, when they are in different classes.
>>>> When using rdfs:label MM creates ID's like #Thing_nr so when
>>>> individuals
>>>> with a same name get mapped to different classes they get the same ID.
>>>> In the expression used there is the class defined as a supertype of the
>>>> individual. I would like the ID of this individual composed with this
>>>> supertype classname, not with Thing.
>>>>
>>>> Can this be done?
>>>>
>>>> Thanks
>>>> Roland
>>>>
>>>>
>>>>
>>>> On 12/13/2010 08:05 PM, Martin O'Connor wrote:
>>>>> Do the classes with the same label pre-exist or are you creating them
>>>>> with MappingMaster?
>>>>>
>>>>> In either case, MM will have no way of distinguishing between classes
>>>>> with the same label.
>>>>>
>>>>> The language does have the option of constructing compound labels and
>>>>> IDs, e.g.,
>>>>>
>>>>> Class: @C4(rdfs:label=(@C4) rdf:ID=("fred_", @C4))
>>>>>
>>>>> which may allow you to partition the names in the expressions. The
>>>>> problem here, however, is that C4 may have characters (e.g., spaces)
>>>>> that cannot be used in an rdf:ID.
>>>>>
>>>>> Martin
>>>>>
>>>>> On 12/13/2010 9:17 AM, Roland Cornelissen wrote:
>>>>>> Hi,
>>>>>>
>>>>>> Using MappingMaster with RDF:LABEL (with RDF:ID auto generated) I run
>>>>>> into trouble when adding instances to different classes with the same
>>>>>> label. Those instances get asserted with different Types and get the
>>>>>> same RDF:ID.
>>>>>> Is there a way to create RDF:ID like in Protege is done when adding a
>>>>>> new instance where the URI gets the name of the class it belongs to
>>>>>> followed by a auto generated ID?
>>>>>>
>>>>>> Now every individual is added as a #Thing; I would like it to be
>>>>>> added
>>>>>> as a #Class. Can this be done?
>>>>>>
>>>>>> Thanks
>>>>>> Roland
>>>>>>
>>>>>>
>>>>>>
>>>>> _______________________________________________
>>>>> protege-owl mailing list
>>>>> protege-owl at lists.stanford.edu
>>>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>>>>
>>>>> Instructions for unsubscribing:
>>>>> http://protege.stanford.edu/doc/faq.html#01a.03
>>>>>
>>> _______________________________________________
>>> protege-owl mailing list
>>> protege-owl at lists.stanford.edu
>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>>
>>> Instructions for unsubscribing:
>>> http://protege.stanford.edu/doc/faq.html#01a.03
>>>
>>
> 
> _______________________________________________
> protege-owl mailing list
> protege-owl at lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
> 
> Instructions for unsubscribing:
> http://protege.stanford.edu/doc/faq.html#01a.03
> 





More information about the protege-owl mailing list