ctCount & Distinct

[BasPav]

Alex,

I might be doing something wrong, but I don't know what:

I build the cube EXTENDED;
I have a measure M of type ctCount;
I check the DISTINCT box:

Now in the Total By Rows I see 1(one) for each line (so far this is what I expect). However, in the Total By Columns I also see a 1 (one). Here I would expect the sum of the above one's, e.g. the number of lines or the count

Please, can you explain this to me.

Thanks, Bas

[AlexZencovich]

Hmm. distinct works in same way as other calulations. It is mean it never sum rows/columns as show in grid. Instead of it aggregate values from multidimensional cube in space for one dimension less then in grid.
Like if you have Client on rows and Ware on columns. It works
1. For grid cell it aggregate cube cells by 2 dimensions - Clients and Wares
2. For Ware totals it aggregate cube cells along Ware dimension
3. For Cleint totals it aggregate cube cells along Client dimension
Which dimension is row, which column - it is indifferently.
In case of distinct it just remove similar values from aggregated set and sum that for ctSumma types and count them for ctCount types or do something else for other types.