Hide null members but show always all non null members

[misscognos]

Hi,

I have data like this:

lev 1    lev 2   lev 3   lev 4   lev 5
  a1       a2      a3     null     null
  a1       a2      a3      a4      null
  a1       a2      a3      a4      a5
  b1       b2      b3     null     null



I use this to create a dimension.

I want to do two things:
1. In row 1 and row 4, i want to hide the null/blank categories generated by nulls starting from the "right side" of the row.
2. Up to the level where the data is not null, I want to include the category regardless of whether there is fact data or not.

In other words, all i want to do is hide nulls but always include non nulls.


I do not want to save the categories in this model as they change very frequently- they should be fresh every time the cube is built.

I tried the suppress blank categories on the levels in my dimension. It does requirement 1 above but also does not display members if there is no fact data for the row. There is always include (which violates requirement 1) and include when needed as well. I would have that "include when needed" would do what I am seeing with suppress blanks.

Any ideas on how I can accomplish this? Also, what is the difference between these two options (suppress blanks and include when needed)? When they say "blank source value" and "current data source" in the documentation snippets below, does it refer to fact data?

Documentation:

Suppress blank categories (levels only)
Retains categories in the model, but excludes from cubes those categories with a blank source value.

Include when needed
Retains the category in the model, but excludes it from cubes when it and all its descendants fail to appear in the current data source.If necessary, ancestors of the category are included, regardless of their Inclusion settings.

[misscognos]

I ended up "initializing" everything to zero by creating a dummy union in my fact query with all possible rows in the dimension. In combination with suppress blanks it does what I need now.