Re: How to solve the matrix kind of scenarios with multiple Dimensions and KeyFigures.?.

As the key figure values depend on the combination of dimension values you can of course calculate all key figure values for a given combination of dimension values in one go.

So that's easy.

More effort typically is to get through generating all or most of the combinations of dimensions.

Often you want to keep one or more dimensions fixed and see the key figure values for the remaining dimensions.

For stuff like this, SQL provides the GROUP BY GROUPS or BY CUBE option (check the documentation for the SELECT command):

create column table multidatefacts (eff_date date, plan_date date, some_date date, any_date date, kf1 decimal (10,2), kf2 bigint, kf3 boolean);

insert into multidatefacts (

select add_days (current_date, gen.rno) as eff_date,

       add_days (current_date, mod (gen.rno, 10)) as plan_date,

       add_days (current_date, mod (rand()*100, 10)) as some_date,

       add_days (current_date, mod (rand()*100, 3)) as some_date,

       to_decimal ((rand()*10000), 10, 2) as kf1,

       to_bigint  ((rand()*1000000)) as kf2,

       to_boolean (mod(gen.rno,2)) as kf3

from ( select top 100000 row_number() over () as rno

    from objects cross join objects) gen);

select eff_date, plan_date, some_date, any_date,   sum (kf1) as sum_kf1, sum (kf2) as sum_kf2

from multidatefacts

group by grouping sets LIMIT 5 WITH TOTAL

     (eff_date, plan_date, some_date, any_date);