- From: Andy Seaborne <andy.seaborne@talis.com>
- Date: Tue, 16 Mar 2010 14:01:01 +0000
- To: Steve Harris <steve.harris@garlik.com>
- CC: SPARQL Working Group <public-rdf-dawg@w3.org>
** Work-in-progress ** I'll try to sort these out in time for F2F3. On 16/03/2010 1:47 PM, Steve Harris wrote: > > One thing, all aggregates are supposed to be able to have DISTINCT as a > flag, I don't think this is possible with the draft grammar. Will do. It's going to mess the function call syntax up though. > > Also, for completeness, maybe we should have ALL, as the opposite, which > is the default (as per SQL)? IMHO I don't see why because the default can't be reset anywhere else. Mostly harmless though. > What about user-defined aggregates too? Can't see them. They are there - regular function calls. Is [] syntax in or out? Need to add [] and DISTINCT > It seems a bit odd that COUNT() only takes Var, not Expression. > COUNT(expr) is potentially useful, eg. Will add. > COUNT(?x && ?y) # count rows where both ?x and ?y are defined. I strongly support this to be the case - my proposed definition of the value multiset to pass the aggregation supports this. Eval failure as aggregate failure does not. And it's COUNT(bound(?x) && bound(?y)) anyway. > - Steve Andy Andy
Received on Tuesday, 16 March 2010 14:01:34 UTC