Gross conceptual question re: JOIN cardinality...

[FerdH4]

I'm just starting to learn Report Studio's JOIN function.  Though I am quite certainly not an SQL expert, I am already (relatively) comfortable with joining tables within SQL. 

My perception is that SQL only asks an author to define a single cardinality; so, I'm struggling with Report Studio JOINS because of what appears to be the requirement to define TWO sets of cardinality values - one for each the primary-to-secondary queries AND also the relationship of the secondary-to-primary queries.

If my perception as stated above is in fact true, can anyone please help explain to me conceptually why Cognos Report Studio demands this two-part cardinality definition?

Is there actually any difference conceptually between these primary and secondary cardinality settings...for example, is one of these two settings seriously more important than the other?

Thanks All.

[Reinhard]

The two cardinalities are just a more detailled depiction of the join, meaning the left one shows the cardinality from the left query subject to the right and the right cardinality the other way round.

In joins between dimensions you only need 1:1 and 1:1 (only exception is an outer join like 0:1).

In joins between dimension and fact table you need 1:N and 1:1 because Cognos identifiers fact tables by the N on the cardinality end.

Regards,
Reinhard

[FerdH4]

Oh, "dimension and fact table".

Thank you Reinhard.