I'm considering the following three ways of partitioning our cube into
months (months of activity):

1. Filter on the date - Example: Date >= '1/1/05' and Date < '2/1/05'
2. Slice on the date - Example: Date.2005.Jan-05
3. No filter or slice: Restrict the date with a view over the fact table
(fact table is not partitioned)

What are the pros and cons are doing both? What would you guys recommend?

Thanks in advance!

to create a parition in a cube, you have to select the target slice (which
is a month) + an SQL where statement (or the usage of another table from the
database instead-of the fact table)

The point 1 you describe is the "where" SQL statement to filter the source
table during the partition process.
The point 2 is the target slice of the cube.
so you have to use the point 1 & 2 in order to create your partition.


"Jesse O" <jesperzz (AT) hotmail (DOT) com> wrote

got it!


"Jéjé" <willgart (AT) BBBhotmailAAA (DOT) com> wrote