Re: [Issue 170] Update on EXISTS

On 11/12/16 13:27, Andy Seaborne wrote:
> The EXISTS CG has two proposals under discussion for fixes to the EXISTS
> feature of SPARQL 1.1
>
> Both address SHACL ISSUE-170 in the same way (by different means). Blank
> nodes passed into EXISTS do not participate in basic graph pattern (BGP)
> nor property path (PP) matching, values of variables are restricted to
> the current row after BGP or PP matching.
>
> ISSUE-68 is not specifically about EXISTS but a revised mechanism for
> EXIST may be applicable for what is current called "pre-binding".
>
> Proposal-A has the simpler definition:
> https://lists.w3.org/Archives/Public/public-sparql-exists/2016Sep/0012.html
>
> Proposal-B covers more uses of EXISTS:
>
> https://lists.w3.org/Archives/Public/public-sparql-exists/2016Sep/0024.html
>
>    and more fully in:
>
> https://w3c.github.io/sparql-exists/docs/sparql-exists.html#an-alternative-to-substitution
>
>
> Any input from the SHACL-WG would be appreciated.
>
>     Andy

If anyone has examples (a few, one or two ...) I can try to run them for 
either proposal-A or proposal-B.  For simple cases, they are often the 
same, it is when there compound patterns (when for example GRAPH, UNION 
and subqueries are used with FILTER, that they differ significantly).

I have a prototype implementation of proposal B. Proposal A is a bigger 
implementation change but I have a partial simulation except for some 
error cases of A (they are not in the linked message - they appear 
elsewhere in the CG email list).

     Andy

Received on Sunday, 11 December 2016 13:59:38 UTC