Syntax.syntaxARQ does not require such naming (well, it invents one that
is an (illegal, to avoid classhes) varname like ?.1).
Sending the results in a standard format with proper names does require
a name of some kind and, if you had multiple aggregates, then there is a
minor issue of how the app finds out the name generated for the aggregate.
But, as noted on the Dryad blog, it's all a bit of an inconvenience for
some usages like poking around in unknown data.
Andy
On 24/09/11 16:36, Tim Harsch wrote:
Oh good. I was hoping there was just something I wasn't seeing. I didn't know
that SPARQL 1.1 required aggregates with no alias.
Thanks Damian!!
----- Original Message -----
From: Damian Steer<[email protected]>
To: [email protected]; Tim Harsch<[email protected]>
Cc:
Sent: Friday, September 23, 2011 5:09 PM
Subject: Re: Syntax.SPARQL_11 doesn't recognize COUNT
On 24 Sep 2011, at 00:04, Tim Harsch wrote:
I'm trying to create some code that will tell me if a query is SPARQL
1.1 or SPARQL 1.0 by first parsing with the 1.0 parser and failing that parse
with the 1.1 parser. But the following fails:
QueryFactory.create("SELECT COUNT(*) {}",Syntax.syntaxSPARQL_11);
com.hp.hpl.jena.query.QueryParseException: Encountered "
"count" "COUNT "" at line 1, column 8.
Was expecting one of:
<VAR1> ...
<VAR2> ...
"distinct" ...
"reduced" ...
"(" ...
"*" ...
at
com.hp.hpl.jena.sparql.lang.ParserSPARQL11.perform(ParserSPARQL11.java:87)
<...SNIP...>
Oops, I read your emails in the wrong order :-)
You need (COUNT(*) AS ?count).
Damian
