MP is works as union operator it brings all records irrespective of the target for the given selection. But the infoset brings data which are common in the infoproviders. Both were structures.
Infosets in BI can join cube also.
Multiprovider Scenario :
infoprovider 1: 0CALMONTH 0PLANT STOCK
infoprovider 2: 0CALMONTH 0PLANT 0VENDOR ORDER_QUANTITY
Some sample data:
infoprovider 1:
201201 PLANT_1 100
201202 PLANT_1 110
infoprovider 2:
201201 PLANT_1 VENDOR_A 70
201202 PLANT_1 VENDOR_B 60
From a technical perspective, you could actually design a MultiProvider having all the 3 characteristics: 0CALMONTH, 0PLANT, and 0VENDOR. But as a matter of fact, it'd be useless.
MultiProvider:
0CALMONTH 0PLANT 0VENDOR STOCK ORDER_QUANTITY
Union of the data in the MultiProvider:
201201 PLANT_1 VENDOR_A 0 70
201201 PLANT_1 # 100 0
201202 PLANT_1 VENDOR_B 0 60
201202 PLANT_1 # 110 0
As you see, the stock data are thoroughly assigned to the vendor code # (i.e. not assigned). That's expected, as the stock data (in infoprovider 1) are not at the vendor granularity level. On the other hand, the order data have the vendor detail.
An InfoSet with the right join condition would present these data:
201201 PLANT_1 VENDOR_A 100 70
201202 PLANT_1 VENDOR_B 110 60
