Re: test cases for fromUnionQuery, please

On Thu, Jun 02, 2005 at 01:48:58AM -0400, Eric Prud'hommeaux wrote:
> Recording Yoshio and my whiteboard discussion here:
> 
> Given this data:
>   <c>:
> 	<A> knows <B>
> 
>   <d>:
> 	<B> mbox <mailto:b>
> 
> It seems the following 4 expressions are all equivilent:
> 
> (1) (Named Graphs (arbitrary pre-existing data)):
>   FROM <c> <d>
>   WHERE {
> 	GRAPH ?g { ?who knows ?whom } .
> 	GRAPH ?g { ?whom mbox ?mbox } .
> 	FILTER ?g = <c> || ?g = <d>
> 
> (2) (Source (arbitrary pre-existing data)):
>   FROM <c> <d>
>   WHERE {
> 	SOURCE ?g { ?who knows ?whom } .
> 	SOURCE ?g { ?whom mbox ?mbox } .
> 	FILTER ?g = <c> || ?g = <d>
> 
> (3) (No Default Graph): 
>   FROM <c> <d>
>   WHERE {
> 	?who knows ?whom .
> 	?whom mbox ?mbox } }
> 
> (?) (Source + CLEAR (arbitrary pre-existing data)):
>   CLEAR
>   FROM <c> <d>
>   WHERE {
> 	?who knows ?whom .
> 	?whom mbox ?mbox } }

I dont believe so, wouldn't

FROM <c> <d>
WHERE {
     ?who knows ?whom .
     ?whom mbox ?mbox }

Be equivalent to

FROM <c> <d>
WHERE {
   SOURCE ?g1 { ?who knows ?whom } .
   SOURCE ?g2 { ?whom mbox ?mbox } . 
   FILTER ?g1 = <c> || ?g1 = <d> &&
          ?g2 = <c> || ?g2 = <d> }

- Steve