Goal: Reduce cube processing time.

Scenario: Dimension has two levels: 1) COST_SET and 2) CLASS. COST_SETs can
have one or more CLASSes associated with them. FACT table has 8 million
rows, each associated with only one class in the CLASS table. Cross
reference table exists to join COST_SET and CLASS tables in a many to many
relationship.

Problem: Rows processed is on the order of FACT table row count * COST_SET
table row count. Even if I set custom aggregations and only aggregate on
CLASS level and not COST_SET, the processing time is the same. I would like
reduce the rows processed to be on the order of the FACT table row count. I
tried dropping the COST_SET level and then creating a virtual dimension that
would add the COST_SET level, but the dimension wizard and editor do not
allow me to add an additional table (COST_SET) to the virtual dimension. Any
suggestions?

Thanks, Nate.

Snowflake or schema has no impact on partition processing if the dimension
has had the schema optimization wizard ran on it. Read the section on
optimizing processing in the SSAS Performance Guide:
http://www.microsoft.com/technet/pro.../ansvcspg.mspx
in the section titled: "Eliminate Joins between Fact and Dimension Tables"
--
Dave Wickert [MSFT]
dwickert (AT) online (DOT) microsoft.com
Program Manager
BI SystemsTeam
SQL BI Product Unit (Analysis Services)
--
This posting is provided "AS IS" with no warranties, and confers no rights.

"Nate" <anonymous (AT) discussions (DOT) microsoft.com> wrote