Re: three hundred dimentions?

Tech-Archive recommends: Repair Windows Errors & Optimize Windows Performance

Thank you very much, Phil.

Let's say we build a fact table with columns like:

OrderDateTime DATETIME,
StoreID INT,
OrderID INT,
ProductItemID INT,
PartOfCombo BIT,
ItemAmount INT

I still am not clear how to build cudes and queries to get the files
results. More help please!

Many many thanks,

DNG


"SQL McOLAP" <SQLMcOLAP@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:C14FE6D5-F457-492D-9F2A-19C0A91A038A@xxxxxxxxxxxxxxxx
> Sorry, I left one aspect out of my last post. It sounds like each line
> item
> (transaction id/product id) in your fact table/view should have a combo
> qualifier (like a 'Y' if it is or an 'N' if it's not) so you can do your
> analysis of whether or not it was included in a combo. If that's included
> in
> your product dimension, then you wouldn't have to worry.
>
> For example, if you had a "hamburger" and "hamburger/part of combo" as two
> different product members.
>
> Another way to do it would be to have a combo dimension that indicates
> which
> combo the product is part of, if any. If the product on that line doesn't
> belong to a combo, it can have the "not part of combo" member assigned to
> it
> from that combo dimension.
>
> Good luck!
>
> - Phil
>
>
>
>
> "DNG" wrote:
>
>> Fast food chain store (Think the likes of Burger King or McDonald's).
>> Items
>> are sold separately or combined. Need to know, for each item sold, what
>> are
>> the ratios that any other items are sold in the same transaction.
>>
>> For example, one day during 5-6PM, 200 Cokes were sold in 160
>> transactions
>> (orders). 75 of them were part of the combo's. Within those 160
>> transactions, there also sold 140 cheese burgers, 130 French fries (85 of
>> them were part of the combo's), . That is, in that time period, 200 Cokes
>> were sold in transaction that also sold 85 French fries in combo, 45
>> French
>> fries not in combo; 140 cheese burgers, etc. The data warehouse is to
>> provide information like: at 8-9AM, 234 Fries were sold in transaction
>> that
>> also sold 48 Cokes and 93 Coffee; at 7-8PM 435 Fries were sold in
>> transactions that also have 238 Cokes and 49 Diet Coke, and so on.
>> Business
>> calendar, store structure, and order size (how much per order) are other
>> dimensions.
>>
>> There are hundreds of items in tens of categories. So, what the best way
>> to
>> design this?
>>
>> No answer from my previous post, so I am going to have a three hundred
>> dimention data warehouse?
>>
>> Thanks a lot!
>>
>> DNG
>>
>>
>>
>>


.

Relevant Pages

  • RE: three hundred dimentions?
    ... product dimension, and a date/time dimension, that drills down to the hour, ... Also, the order ID distinct count measure should be in a cube of its own, ... Need to know, for each item sold, what are ... > the ratios that any other items are sold in the same transaction. ...
    (microsoft.public.sqlserver.olap)
  • RE: three hundred dimentions?
    ... your product dimension, then you wouldn't have to worry. ... Need to know, for each item sold, what are ... > the ratios that any other items are sold in the same transaction. ... > were sold in transaction that also sold 85 French fries in combo, ...
    (microsoft.public.sqlserver.olap)
  • A challenge design question
    ... Fast food chain store. ... Need to know, for each item sold, what are ... were sold in transaction that also sold 85 French fries in combo, ... StoreID INT, ...
    (microsoft.public.sqlserver.olap)
  • Re: Transactional Replication Problem
    ... commit transaction ... goto EndSave ... execute msdb.dbo.sp_add_category N'REPL-Distribution Cleanup' ...
    (microsoft.public.sqlserver.replication)
  • Re: About MSDTC,The same ContextUtil.TransactionId,but the state of Database A,B is different.
    ... Transaction(TransactionOption.Required)] ... > I think the only one MSDTC is used,MSDTC ... > int CommitTransaction(); ... > connectionstring,string SqlStr,out string ErrStr) ...
    (microsoft.public.dotnet.framework.adonet)