- From: Andy Seaborne <andy@apache.org>
- Date: Thu, 29 Sep 2016 00:44:32 +0100
- To: public-sparql-exists@w3.org
Jeremy - great example.
I think it is in-scope.
Andy
On 27/09/16 17:19, Jeremy J Carroll wrote:
> james andersion suggested:
>> it would be valuable to reduce the query and the dataset to the minimum case which demonstrates the deficiency,
>> record the correct result, and include the case in the test suite.
>
> I found this case to be simple, with an empty dataset.
>
>
> select *
> { BIND (1 as ?x)
> FILTER EXISTS {
> # Uncomment optional block to correct
> # OPTIONAL {
> # <http://does.not.exist> <http://does.not.exist> ?x
> # }
> FILTER EXISTS {
> FILTER (?x = 1)
> }
> }
> }
>
>
> correct answer is a single binding of ?x with value 1.
> The particular implementation I am using gets confused with a doubly nested FILTER EXISTS: a confusion which the OPTIONAL block referring to ?x resolves.
>
> Jeremy
>
>
Received on Wednesday, 28 September 2016 23:45:03 UTC