Re: Date Concepts

On Thu, 7 Nov 1996 05:11:16 GMT, rdshaver_at_tscnet.com (Rick and Debbie Shaver) wrote:

>So how do you resolve this?? I'm sure we aren't the only ones that
>have crossed this bridge. Do you go back and correct the 11/1 time
>slice?? (I understand that's taboo).

You have to, otherwise your Membership figures are incorrect !

It is quite easy to model this with a single Time dimension (Membership Date). But, with the one Time dimension your warehouse is going to be able to answer questions such as "Membership by Month, as at any point in time". It is possible to introduce a second Time dimension (As at Date), but we chose not to, as data volumes would be too large.  

>
>I also have another question on this time slice thing. Say I have
>table that contains 12 months of summarized claims payments for each
>doctor. Each record has 12 buckets, one for each month of the year.
>(I understand this is the kind of thing you want in a DW). Each month
>you plop the new summarized amount into the appropriate bucket. So in
>the dw time slice for Jan you capture the Jan bucket. On Jan 15 (the
>second time slice) do you capture the Jan bucket again?? Then you
>have redundant data if you try and do any trending analysis through
>the time slices. What about on Feb 1st, do you capture the Jan and
>the Feb bucket??

The Claims problem is very easy. You will need two Time dimensions. A Treatment Date and a Processing Date. There is no redundant data this way.

Good luck !

Mike Adendorff
madnrf_at_iafrica.com Received on Sun Nov 10 1996 - 00:00:00 CET