I'm trying to find the Pearson correlation of a single measure between two members of a dimension, but the msdn docs a开发者_如何学JAVAre a bit sparse. Specifically I have a cube with a fact count
measure, a date
dimension, and a tool
dimension and I'd like to find the correlation of tool X and tool Y over the date dimension.
take a look at the following script (Adventure Works DW 2008 R2):
It will return correlation of [Internet Sales Amount] measure for two different product subcategories ("Mountain Bikes"/"RoadBikes") for months of current date member on rows (Calendar Year 2007 quarters and Calendar Year 2007). I have left other comparable members in comments.
with
member ActualMeasure AS [Measures].[Internet Sales Amount]
member m1 AS
(
[Product].[Product Categories].[Subcategory].&[1] -- Mountain Bikes
-- [Sales Territory].[Sales Territory].[Group].&[North America]
-- [Customer].[Gender].&[F]
,ActualMeasure
)
member m2 AS
(
[Product].[Product Categories].[Subcategory].&[2] -- Road Bikes
-- [Sales Territory].[Sales Territory].[Group].&[Europe]
-- [Customer].[Gender].&[M]
, ActualMeasure
)
member x as
Correlation
(
{Descendants([Date].[Calendar].CurrentMember,[Date].[Calendar].[Month]) } as dates
, m1
, m2
), Format_String="Standard"
select
{ x,m1,m2 } on 0,
{
Descendants
(
[Date].[Calendar].[Calendar Year].&[2007]
, [Date].[Calendar].[Calendar Quarter]
)
,[Date].[Calendar].[Calendar Year].&[2007]
} on 1
from [Adventure Works]
HTH,
Hrvoje Piasevoli
精彩评论