MDX Query with running total across crossjoined dimensions - ssas

Question

I have a cube with the following three dimensions: Date, Time and Shift. I have a measure called [Pieces Succeeded], and I want a running total of the [Pieces Succeeded] by hour for a Shift. A Shift can span more than one day, so in the following query, I do a crossjoin of the Date and Time dimensions.
with
member [Measures].[Pieces Succeeded Running Total] as
sum([Time].[Hierarchy].[Hour].FirstMember:[Time].[Hour].CurrentMember, [Measures].[Pieces Succeeded])
select
{ [Measures].[Pieces Succeeded], [Measures].[Pieces Succeeded Running Total] } on columns,
nonempty(crossjoin([Date].[Month Hierarchy].[Day].Members, [Time].[Hierarchy].[Hour].Members)) on rows
from
[OEE]
where
[Shift].[Month Hierarchy].[Shift].&[501]
Which gives the following results:
Date Hour Pieces Succeeded Pieces Succeeded Running Total
03 Apr 2011 22 6393 6393
03 Apr 2011 23 6424 12817
04 Apr 2011 00 3816 3816
04 Apr 2011 01 5510 9326
04 Apr 2011 02 2090 11416
04 Apr 2011 03 7489 18905
04 Apr 2011 04 7307 26212
04 Apr 2011 05 5706 31918
How would I go about getting the sum to work on the crossjoined set so that the Running Total works across days?
Thanks

Answer 1

I spent all day on this, and finally figured it out. I thought it might be valuable for someone else, so here's the solution:
with
set DateHours as
nonempty(crossjoin([Date].[Month Hierarchy].[Day].Members, [Time].[Hierarchy].[Hour].Members), [Measures].[Pieces Succeeded])
member [Measures].[Rank] as
rank(([Date].[Month Hierarchy].CurrentMember, [Time].[Hierarchy].CurrentMember ), DateHours)
member [Measures].[Running Pieces Succeeded] as
iif([Measures].[Rank] = 1, [Measures].[Pieces Succeeded], sum(head(DateHours, [Measures].[rank]), [Measures].[Pieces Succeeded]))
select
{ [Measures].[Pieces Succeeded], [Measures].[Running Pieces Succeeded] } on columns,
non empty { DateHours } on rows
from
[OEE]
where
[Shift].[Month Hierarchy].[Shift].&[501]