![]() | |
![]() |
| | Thread Tools | Display Modes |
#1
| |||
| |||
|
#2
| |||
| |||
|
|
Hi, In our old MSAS 2000 we have about 20 physical cubes (20 fact tables and around 300 rollup tables). Everything is created in 1 data base. Some of the physicsl cubes share dimensions (rollup tables). When we migrated to MSAS 2005 with the help of Migration wizard, it created 1 huge DSV with 20 fact tables and 300 rollup tables. As you can understand the DSV is unmanageable and very hard to interpret. Please suggest any alternative designs. One thing i did is to break the one huge DSV into multiple logical smaller DSV's to easily manage. the problem is once i start building physical cubes from these different DSV's, each cube created its own copy of the dimension as the source DSV is different. Now that i have 5 logical DSV's and some common shared dimensions, i have multiple copies of the same dimensions with different names. More over the dimension names it created are meaningless. Please comment. Thanks, Sam. |
#3
| |||
| |||
|
|
in 1 DSV you can create "views" (not SQL views) in these "views" you can display only the table you want. the "default view" contains all the tables. lookat the upper left of the DSV screen to create a new one. "Prasad" <Prasad (AT) discussions (DOT) microsoft.com> wrote in message news:128A49B5-D388-43B5-AE34-4DE319BF0CF0 (AT) microsoft (DOT) com... Hi, In our old MSAS 2000 we have about 20 physical cubes (20 fact tables and around 300 rollup tables). Everything is created in 1 data base. Some of the physicsl cubes share dimensions (rollup tables). When we migrated to MSAS 2005 with the help of Migration wizard, it created 1 huge DSV with 20 fact tables and 300 rollup tables. As you can understand the DSV is unmanageable and very hard to interpret. Please suggest any alternative designs. One thing i did is to break the one huge DSV into multiple logical smaller DSV's to easily manage. the problem is once i start building physical cubes from these different DSV's, each cube created its own copy of the dimension as the source DSV is different. Now that i have 5 logical DSV's and some common shared dimensions, i have multiple copies of the same dimensions with different names. More over the dimension names it created are meaningless. Please comment. Thanks, Sam. |
#4
| |||
| |||
|
|
in 1 DSV you can create "views" (not SQL views) in these "views" you can display only the table you want. the "default view" contains all the tables. lookat the upper left of the DSV screen to create a new one. "Prasad" <Prasad (AT) discussions (DOT) microsoft.com> wrote in message news:128A49B5-D388-43B5-AE34-4DE319BF0CF0 (AT) microsoft (DOT) com... Hi, In our old MSAS 2000 we have about 20 physical cubes (20 fact tables and around 300 rollup tables). Everything is created in 1 data base. Some of the physicsl cubes share dimensions (rollup tables). When we migrated to MSAS 2005 with the help of Migration wizard, it created 1 huge DSV with 20 fact tables and 300 rollup tables. As you can understand the DSV is unmanageable and very hard to interpret. Please suggest any alternative designs. One thing i did is to break the one huge DSV into multiple logical smaller DSV's to easily manage. the problem is once i start building physical cubes from these different DSV's, each cube created its own copy of the dimension as the source DSV is different. Now that i have 5 logical DSV's and some common shared dimensions, i have multiple copies of the same dimensions with different names. More over the dimension names it created are meaningless. Please comment. Thanks, Sam. |
![]() |
| Thread Tools | |
| Display Modes | |
| |