Seemingly simple MDX calculation but using the same dimension twice with different variables

Question

Table:

X             Y             wins  losses
Apples        Oranges       5     2
Apples        Limes         2     1 
Apples        Strawberries  3     3
Limes         Pears         5     4
Limes         Strawberries  6     8
Strawberries  Limes         2     3
Strawberries  Apples        1     5

I need to sum [X].Apples wins where Y = Strawberries and divide that by [X].Strawberries total losses. [Measures].[some_calculated_member] = 3 / (5 + 3)

An alternative would be to calculate it this way: SUM([X].Apples wins where Y = Strawberries / SUM([Y].Strawberries wins)

I'm not sure how to structure the MDX query to accomplish what I need:

SELECT
{ [Measures].[some_calculated_member] } ON 0,
{ [X].[Apples] } ON 1,
{ [X].[Strawberries] } ON 2,
{ [Y].[Strawberries] } ON 3
FROM [cube]

I've looked into using AXIS but I'm using Mondrian, which doesn't support AXIS.

Answer 1

Did you try with a calculated measure?

WITH 
Member [Measures].[Ratio] AS (
  [Measures].[wins.SUM],
  [X].[ALL].[AllMember].[Apples],
  [Y].[ALL].[AllMember].[Strawberries]
) / (
  [Measures].[losses.SUM],
  [Y].[ALL].[AllMember].[Strawberries]
) 
SELECT NON EMPTY {
  [Measures].[New Measure]
} ON COLUMNS 
FROM [SomeCube]