Re: Multi Dimensional Cases in Time Series

In case if you are interested, I blogged the problem and the solution at
http://solidqualitylearning.com/blogs/Dejan/archive/2006/01/27/1512.aspx.

--
Dejan Sarka, SQL Server MVP
Mentor, www.SolidQualityLearning.com
Anything written in this message represents solely the point of view of the
sender.
This message does not imply endorsement from Solid Quality Learning, and it
does not represent the point of view of Solid Quality Learning or any other
person, company or institution mentioned in this message

<Panorama@xxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:bd6e429e-ec48-47d3-9655-fa7c9de473df@xxxxxxxxxxxxxxxxxxxxxxx
> Thanks, Dejan!
>
>
>
>>
>> > Is there any way to have multiple dimensions as cases for the time
>> > series
>> > algorithm?
>> > For example, you can have both "Customer" and "Product" dimensions as
>> > cases for the time series algorithm, so you'd get, for example, the
>> > time
>> > series results of the customers in "Washington" buying product category
>> > "Food".
>>
>> I don't think it is possible directly. Data Mining structure supports
>> only a
>> single level of nesting tables, and this is level used by a fact table,
>> no
>> matter from which dimension you start.
>> The best solution I can remember at the moment is:
>> - in SQL Server, create a linked server to the Analysis Services
>> - in SQL Server, create a view by using Qpenquery T-SQL function
>> - in the body of the Openquery use MDX query to retrieve the data from
>> the
>> cube
>> - take care to return a two-dimensional set only!
>> - mine the view.
>>
>> --
>> Dejan Sarka, SQL Server MVP
>> Mentor, www.SolidQualityLearning.com
>> Anything written in this message represents solely the point of view of
>> the
>> sender.
>> This message does not imply endorsement from Solid Quality Learning, and
>> it
>> does not represent the point of view of Solid Quality Learning or any
>> other
>> person, company or institution mentioned in this message
>>
>>


.