I read alot of your posts and they are really awesome and your posts really helped me in some tricky situations. i have one issue presently while designing the universe i have two fact tables say A and B and i have 5 conformed dimension tables and additional 3 dimension tables which are joined to B. i have a senario where i need to use a measure from B via A at present we have a join between two fact tables and there is only one measure in B. I created two contexts for A and B at present for now performance is degraded because i have a join between two fact tables(both are huge) how can i get the measure from B via A without involving the join between the two fact tables?

Do you have any good example to Understand the Aggregate Awareness and Aggregate Navigation, this will be helpfull.

Thxs,

Thanks for the reply. I’m still going through your blogs… currently i’m on JOIN_BY_SQL….. really interesting.
I’m just done with my office work.
Your blogs are not only interesting but the way you’ve narrated them they become really catchy.

When can we see your next post on chasm trap… and also i want to know which joins i should select in a context?

Thanks again.

Regards,
Shiva

so as per the above statement….in any M:1:M OR M:M condition i need to use both(mandatorily) to resolve the issue completely. Is it true?

Many to One to Many is a Chasm trap, which is easily resolved with two contexts, one for each side of the relationship.

“Many to Many” logical relationships are generally implemented with an extra table so you end up with 1:M:1. The “M” table becomes the “fact” table in this case and requires a context.

No aliases are required in either of those situations.

Shortcut joins are not a loop resolution mechanism. They are used to skip tables for performance reasons only. Shortcuts should not be added to a universe until everything else is verified and working.

How do i identify when to use a Alias and when to use a context.

As per your blog above
“An alias was indicated by tables on the “one” side of every join. Contexts are indicated by tables appearing only on the “many” side of a one-to-many join./”

so as per the above statement….in any M:1:M OR M:M condition i need to use both(mandatorily) to resolve the issue completely. Is it true?

But if i break a loop using Aliases then why do i need a context?

(I know you’ll ask me now .. what if i need data from tables wherein one table is at the begining and the other is at the end of the joining path?)

Shortcut join can resolve this(I guess)
but that would form a loop again? For which again we’ll need a context.
As we need to tell BO that i need this context(joining path) to get my data?

This in short means we always need both (aliases and contexts) to resolve a M:M cardinality completely??

Please Clarify.

Thanks a bunch.
Shiva

The primary issue with unresolved loops is that Business Objects will generate SQL that takes every available path at the same time. Thus, you will get SQL that uses both paths (service + invoice).

It’s easier to explain with the original COUNTRY example. Each customer has a country. Each resort has a country. There is a loop (not shown in this post) from Country to Region to Customer to … to Resort and back to Country again. If that loop is not resolved, the entire path is used for every query. The result? You will only get rows where a customer visits a resort in the same country where they live, as otherwise the join path will fail.

Could you please give more examples of the loops.
Will it be a loop if the relationships are (C 1:M R1 1:M S) and (C 1:M R2 1:M S)?
Will it be a loop if the relationships R1 and R2 use different alternative keys in S?

Thanks.