I have been taught that using outer joins is a no-no in the world of Cognos data-modelling.
However, I need to rebuild some BO universes that work with outer joins.
These universes relate to OLTP schema's and some outer joins cannot possibly be eliminated.
Now, there is no option to use a DWH (cause this is a temporary situation)
The second issue is that of factless facttables. Some transaction tables have zero facts.
In BO modelling I would create measures like:
and this measure would be aggregated against dimensions.
pure and simple from an SQL perspective.
So, the second question would be: how to create the facts?
Ties Blom
However, I need to rebuild some BO universes that work with outer joins.
These universes relate to OLTP schema's and some outer joins cannot possibly be eliminated.
Now, there is no option to use a DWH (cause this is a temporary situation)
The second issue is that of factless facttables. Some transaction tables have zero facts.
In BO modelling I would create measures like:
Code:
count(object)
or
count(distinct object)
and this measure would be aggregated against dimensions.
pure and simple from an SQL perspective.
So, the second question would be: how to create the facts?
Ties Blom