Re: hamsterdb Transactional Storage (thanks to all of you)

Bob Badour wrote:
> compdb_at_hotmail.com wrote:
>

>> On Oct 5, 10:56 am, rp..._at_pcwin518.campus.tue.nl (rpost) wrote:
>>
>>>> If you have a generalized aggregate corresponding to series notation
>>>> you can express UNNEST as UNION over a set of tuples.
>>
>> If we could define aggregates like SUM(R, X) and PRODUCT(R, X+Y) for
>> arbitrary operators via a generalization of standard math SIGMA and PI
>> series notation over elements of a set then we could define R UNNEST A
>> as the iterated UNION over tuples t in R of expression (relation{t}
>> REMOVE A) JOIN t.A. So generalized aggregates give it for free.

I've been puzzling over this. Is it a way of allowing user-defined aggregate ops? And what "detritus" is avoided? Also does it mean a "tuple" programming interface? Or does it simply mean that most implementations don't define all the possible ops for each domain? I might be able to get the point with an answer to just one of those questions.

(I take it that UNNEST here has a tuple operand, unlike UNGROUP. I've presumed that the conventional aggregate definitions involve UNGROUP as well as EXTEND.) Received on Tue Oct 13 2009 - 16:25:09 CEST