Re: constraints in algebra instead of calculus

From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Mon, 18 Jun 2007 13:23:34 -0300
Message-ID: <4676b17e$0$4341$9a566e8b@news.aliant.net>

David Cressey wrote:

> "paul c" <toledobythesea_at_oohay.ac> wrote in message
> news:MtUci.31107$NV3.16822_at_pd7urf2no...
>

>>paul c wrote:
>>...
>>
>>>...  In other words, when all
>>>attributes are grouped, the result has the same number of rows as the
>>>input.
>>>...
>>
>>If this is right, what I don't see is how one could GROUP a table with
>>more more than one row and get a result that had only one row unless
>>some projection to eliminate an attribute was made after grouping on all
>>attributes except the one that is subsequently projected away.  Not sure
>>if that matters though.

>
>
> PMFJI. I'm not following much of this discussion, but I want to ask the
> following.
>
> If you were to GROUP on no attributes at all, wouldn't you get a result
> with only one row?
> Does "GROUP on no attributes at all" even make sense?

That would extend each tuple with a named attribute equal to DEE. So, yes, the relation value attribute added to each row would have one row. But the cardinality of the derived relation would equal the cardinality of the original relation. Received on Mon Jun 18 2007 - 11:23:34 CDT