Re: changes in XQuery and SPARQL's use of XQuery

Steve Harris wrote:
> On Fri, Nov 18, 2005 at 01:41:20PM +0000, Andy Seaborne wrote:
> 
>>
>>Steve Harris wrote:
>>
>>>On Fri, Nov 18, 2005 at 10:14:22 +0000, Andy Seaborne wrote:
>>>
>>>
>>>>>I would be in favour of allowing stuff like "1" + "2" = 3, it fits with
>>>>>how people really use RDF in my opinion.
>>>>
>>>>On the other hand:
>>>>"1" + "2" = "12"
>>>
>>>
>>>As far as I remember we dont have a + operator that takes two strings and
>>>concatentates them, and if we do, I want my money back ;)
>>
>>No ... I have had requests for more string manipulation functions ... so 
>>"not yet"
> 
> 
> I have no problems with a concat type function (concat("a", "b") == "ab"),
> but overloading + is a mistake IMHO. c.f. JavaScript.
>  
> 
>>Actually, + isn't the best example: FILTER(?x < 3) is a better example 
>>where an expectation might be force ?x to integer (c.f. RDQL!)
> 
> 
> Right, which is what I do in my non-standards compliant mode.

So what about something like "1" = 1 ? In non-standards mode, does 3Store 
return true/false/error for this case?

My example of "01" = 1 is more complictaed than this - even when used with 
canonical forms, I'm not clear what to do in the canonical case.

Does this apply when one side is an explicit constant (so the query knows the 
type) and generates a query that is the same as something like:

?x = 1  ==>  xsd:double(?x) = 1

	Andy


>  
> 
>>But then you get the RDQL problems:
>>
>>What about:  "01" = 1  ?
> 
> 
> Sure, its no panacea.
> 
> - Steve
>

Received on Wednesday, 23 November 2005 17:00:13 UTC