April 1, 2013 at 11:26 pm
Comments posted to this topic are about the item Low Profile Architecture
April 2, 2013 at 1:50 pm
Two things immediately spring to mind.
1) What peer review process is undertaken prior to articles like this being published?
2) Was it meant to be published yesterday (April 1)?
...One of the symptoms of an approaching nervous breakdown is the belief that ones work is terribly important.... Bertrand Russell
April 2, 2013 at 2:49 pm
Could you add some specific here?
April 3, 2013 at 10:56 pm
@gpo I think you make some very good points - who peer reviewed this?
I think there is some value in the article but by no stretch of the imagination would you call this a data warehouse or even a data mart.
If the objective is to count unique transactions per client per month.... why not use COUNT(DISTINCT) in your SQL?
And if you are summarising transactions, you have to pick one comment regardless.
I don't mean to be picky but there are quite a few improvements to be made here.
April 4, 2013 at 7:18 am
Thank you for your comment. It gives me an opportunity to explain missed parts.
My goals was:
1. Provide necessary monthly counts (measures).
2. Be able to present detailed records (consider my Comments field as number of joins to tables providing payment, demographic and so on).
3. Minimize ETL related objects (tables and scripts).
Count(distinct *) was an initial solution but as any aggregate it brings difficulty to provide detail records.
As for DW - Excel pivot tables are one of the way to present a multidimensional structures.
So all I need for virtual DW to prepare dimension tables and source to count measures.
Main problem on my way were the duplicates and monthly granularity.
I concentrated in my article on this task for simplicity but seems it makes it vague.
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply