Hi,

I need to aggregate absolute value on each dimension member on a dimemsnion
level. Here is the sample data.

L1 = 200
L1 = 400
L1 = -300
L2 = 100
L2 = -300
......

The standard SUM aggregation issue here is 100 (200+400-300+100-300). But I
don't want to sum them up directly. Instead, I need to sum absolute value of
L1 and L2 individually. The right calculation is below and the value should
be 500.

L1 = 200 + 400 - 300 = 300
L2 = 100 + (-300) = -200 * -1 = 200

==> L1 + L2 = 500 (300 + 200)

Can someone suggest me how to do that in AS function or in MDX calculated
members? Thanks.

-Charles