Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.misc -> Re: Help with SQL constraint
On 16 fév, 17:24, DA Morgan <damor..._at_psoug.org> wrote:
> hasta..._at_hotmail.com wrote:
> > On 16 fév, 01:48, DA Morgan <damor..._at_psoug.org> wrote:
> >> dean wrote:
> >>> On Feb 15, 11:41 am, DA Morgan <damor..._at_psoug.org> wrote:
> >>>> Arto Viitanen wrote:
> >>>>> dean wrote:
> >>>>>> Hello all,
> >>>>>> A table T has 2 fields, one (L) holding letters 'Y' and 'N', and one
> >>>>>> (X) holding numbers. Is there a (non trigger)constraintsuch that for
> >>>>>> L='N' (and only this letter) the numbers must be unique? Records
> >>>>>> where L='Y' do not have to be unique.
> >>>>>> (I need to join another table to the L='N' group of records, and the
> >>>>>> join must be key-preserved).
> >> So, essentially, you want to NOT have unique data but convince Oracle,
> >> by some hocus-pocus magic trick, that the data is unique.
>> >>helpavailable from me. > YOYO. There is one and only one solution ...
> >> As I presume you intend to put this into production there is no more
>
>> > sensor that detects whether a car is passing through (actually, we
> > You are monitoring N crossroads. On each crossroad, there is a
>
>> > through each crossroad within period of times of - say - one hour.
> > You have to count and keep an history of the number of cars passing
>
>
>
That's indeed the best solution, Daniel.
However, I stressed that traffic is huge. You have may store 2 or 300 times more information than what the requirements ask for.
When the history to be kept is huge by itself, then it makes (at least some) sense to accumulate the data at the acquisition level. Received on Fri Feb 16 2007 - 11:09:47 CST