Re: basic vs group graph pattern

On 30/05/17 20:56, Toni Schindler wrote:
> Hi.
> 
> Given the data
> 
> :x1 :p "s1"
> :x2 :p "s2"
> 
> I expect query 1
> 
> {
>   ?x1 :p ?s1.
>   ?x2 :p ?s2.
> }
> 
> to return two solutions,
> 
> ?x1    ?s1    ?x2    ?s2
> :x1    "s1"    :x2    "s2"
> :x2    "s2"    :x1    "s1"

It has 4 solutions.

?x1 :p ?s1.

can match ?x1=:x1, ?s1=:s1 and also ?x1=:x2 ?s1=:s2

and then

?x2 :p ?s2.

can match ?x2=:x1, ?s2=:s1 and also ?x2=:x2 ?s2=:s2

There is no connection between the two parts so you get 2 x 2 = 4 
solution - every combination.

There is no reason

?x1=:x1, ?s1=:s1, ?x2=:x1, ?s2=:s1

isn't a solution. Just because ?x1 is matched to :x1 does not stop ?x2 
being matched the same way.

 Andy

> 
> while query 2
> 
> {
>   { ?x1 :p ?s1. }
>   { ?x2 :p ?s2. }
> }
> 
> should have four solutions:
> 
> ?x1    ?s1    ?x2    ?s2
> :x1    "s1"    :x2    "s2"
> :x2    "s2"    :x1    "s1"
> :x1    "s1"    :x1    "s1"
> :x2    "s2"    :x2    "s2"
> 
> This is because in the first case, there is one length-2 basic graph
> pattern, matching a length-2 subgraph (the whole graph) with two
> different ways to bind the variables to the matched graph.
> In the second case, there is a group graph pattern of two length-1 basic
> graph patterns, each matching two subgraphs, and in each case there is
> one way to bind variables; both length-2 solutions are joined, giving a
> total of four solutions.
> 
> Can someone confirm that this is the correct interpretation?
> This is a reduced example of my question about a test case I asked here:
> https://lists.w3.org/Archives/Public/public-rdf-tests/2017May/0000.html
> 
> Toni
> 
>

Received on Tuesday, 30 May 2017 21:04:45 UTC