For AS, is it mandatory to use Star/Snowflake schemas as datasource?
I have a single table completely denormalized with the measures and
dimensions already in them with a schema like this:

measure1,measure2,dimension1,dimension2,dimension3 ,dimension4,dimension5

The description about is a column description of the measures and
dimensions. The dimension member names provided in this table are all
leaf-level members. I will manually create the levels. My question
is, can I use this table's schema to read into the cube? If so, how?
What will I use for the member-key-column property?

yes, you can

you can create some local dimensions in a cube
or you can create shared dimensions based on your table

your final cube will use only 1 table without any problem.

"Rico Cali" <ricocali (AT) hotmail (DOT) com> a écrit dans le message de news:
675e1895.0409051900.6178dec2 (AT) po...OT) google.com...

Thank you Jeje, but why does MS recommend a star/snow-flake schema for
data-source? Is it because it's faster to read?

because generally its easy to maintain in most cases.
If you have 1 table with a clean list of customers, its easier to create a
dimension based on this table and use another table for the fact table (the
sales information)
also, if you look at the queries launched by AS, you'll see that AS run
Select DISTINCT statements to fill its dimensions, because each customer
appear only 1 time in a dimension, if you fact table has 1 000 000 rows for
only 100 customers, the distinct command takes a very long time.
also, if the name of the customer is 50 characters * 1 000 000 = 50 000
000bytes huge table !!! but a customer key which is an INTEGER type takes 4
bytes * 1 000 000 = 4 000 000bytes

also, if want to share the same dimension on 2 different cubes, AS will
create inner join between your dimension table and your fact table (except
in specific cases not describe here), so an inner join between 2 tables with
1 millions of row each is very long!

its why a star schema is recommended.

but if your cube is based on a small table, then, there is no problems.

I hope this can help you.

Jerome.


"Rico Cali" <ricocali (AT) hotmail (DOT) com> a écrit dans le message de news:
675e1895.0409060902.5f19e18a (AT) po...OT) google.com...

Thank you Jerome, you have been a big help