TM1 Forum

In CUBE OPTIMIZATION i needed in using CUBEVIEW as data source for TI process, is there is any difference in cube performance if the orded of dimemsion in cube and the order of dimension in cube view of same cube that is being used as data source in TI process is different? if so how much and any suggestion to overcome it???

For example my cube is created with dimensions as:

version
scenario
currency
country
depatment
operating unit
product
year
account
month


And order of dimension in cube view of same cube that is being used as data source in TI process is

country
department
operating unit
version
currency
scenario
product
year
month
account


Please provide your valuable suggestion .

The order of variables in a TI process should ALWAYS be the same as the originally defined cube dimension order. This is CRITICAL functionality for generalized processes used with multiple cubes as sources. However if you search this forum you will see that IBM have managed to break this on occasion.

In 10.1 there is an issue where if a cube view created with the cube viewer is used as a data source then the variable order will be jumbled. This could well be your issue. Make sure any view used for a TI source is created inside the TI wizard or with the right-click "export data" wizard if creating manually. Then you will get the variable order as per the cube dimension order.

Thanks a lot for your reply.

So should i arrange dimensions of CUBE-VIEW in same order as they are arranged in CUBE for better performance??