union (combined Query) vs Merge Dimension

[Former Member]

Hi All,

Issue Description: if i want to do report on full data of two universes (univ 1 (source “a”and univ2 source “b”) how can i do the union of the universe? Web intelligence allows me to join (union , intersection , minus) only on query of the same universe but not on different universe! (correct me if I am wrong) and also when we can get the union data my using merge dimensions what is the use of combined query(union option).
help me is it possible to do this or to have a universe like this?

[sateesh_kumar1 · Accepted solution]

Hi Stefen,

if i want to do report on full data of two universes (univ 1 (source “a”and univ2 source “b”) how can i do the union of the universe?

* Use Merge Dimension with Extended Merge Dimension property enabled.

This will give you merged dimension values extended to all values from both queries.

Web intelligence allows me to join (union , intersection , minus) only on query of the same universe but not on different universe!

Yes .

If you are getting data from different data providers then use Merge dimension to get common data .

[Former Member]

Note that the "set" operations (union, intersection, minus) are not going to give you the same type of results as queries with merged dimensions.

Union, intersection, and minus work with queries that select the same columns (the same objects in your object pane).  Rows from the results are either combined (UNION) or selected (INTERSECT, MINUS...) based on a comparison of the results.  You're comparing apples to apples, if you will -- and deciding which ones to keep to make applesauce.   This is why you can't use these operators with different universes, since there would then be no guarantee that you're comparing compatible results.

Merged dimensions take results from different queries which do not necessarily have the same columns or objects and determine a rule to combine them into one merged result.  The merged dimensions form a "join", if you will.  If the dimensions can be merged, they are compatible.  Measures from both queries can be aggregated (or rather projected by BO) according to these merged dimensions.  Any dimension which can't be merged remains compatible with only the results from the query from which it came, but not the other: otherwise you're comparing apples to oranges.

[former_member182342]

Hi stefen jay,

Can you please let us more details about the report . ex. Backend etc.

Regards,

Anish