Hi all,

Please don't tell me to look in microsoft kb 307276. This
does not resolve anything.

The problem is the functions are not working to the BNF.
Even SQL Server Book online is not reliable to resolve
our issues of missing information.

I am looking for a source of information and valid BNF
syntax about the LinRegxxx functions.

Even more right now the results we get with the
LinRegPoint are not even close to what we should get.

We would like a way to predict the next month using the
previous 12 months.

Is there any hidden functions available to get evaluate
other type of Regression?

Why ChartSpace is able to have many type of regression?

regards,

Guy

Here is an example for the Sales Cube of Foodmart, using LinRegPoint to
project Sales for the next month, based on the trailing 6 months
(selected month included). It can be modified to 12 months, but Sales
Cube only has 1997 data:

Member Measures.SalesTrend as
'LinRegPoint( 7, LastPeriods(6) as Trailing6,
[Measures].[Store Sales],
Rank(Time.CurrentMember, Trailing6) )'

Select
{ [Measures].[Store Sales], Measures.SalesTrend }
on Columns,
{ [Time].[1997].[Q2].[6] :[Time].[1997].[Q4].[11]}
on Rows
from Sales

- Deepak

*** Sent via Developersdex http://www.developersdex.com ***
Don't just participate in USENET...get rewarded for it!