W3C home > Mailing lists > Public > public-sparql-dev@w3.org > April to June 2015

Re: Subject lists in SPARQL

From: james anderson <james@dydra.com>
Date: Tue, 7 Apr 2015 13:37:13 +0000
Message-ID: <0000014c94196129-114cd02a-9302-4816-94e1-fc0bf4829b80-000000@eu-west-1.amazonses.com>
To: public-sparql-dev@w3.org
good afternoon;

On 2015-04-06, at 16:26, Fabiano Luz <fabianocomp@gmail.com> wrote:

> Hi guys the question is as follows.
> I'm working on a model of semantic parsing. I try to correlate a subset of natural language with a SPARQL subset.
> For example, for the phrase: "What Bob's phone number and e-mail?"
> We could represent through SPARQL:
> PREFIX p:<example#>
> SELECT ?A ?B {p:Bob p:phone ?B; p:email ?A}
> Note that there is a relationship between the amount of elements of the sentence and the query, so, is possible mapping in our model.
> Using the same reasoning for the phrase: "Which are Bob's and John's phone number?"
> Here, the ideal is a query that I need not repeat the word "phone" because, in my model, repeat terms implies in trick. In addition, the query needs to return a list of phones (Bob's phone and John's phone) and I can not use FILTER.        

you describe, that you would like to specify the sub-graph in which a given subject set is connected to the respective object terms through given predicates.
in what sense is this not denoted by a values clause combined with a bgp?

best regards, from berlin
james anderson | james@dydra.com | http://dydra.com
Received on Tuesday, 7 April 2015 13:37:45 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 7 April 2015 13:37:45 UTC