Hi,

I have a huge datawarehouse (atleast 300 million rows). I have seperate date
and time dimensions(hour and minute attributes) to reduce dimension size.
However most of our MDX queries are date range queries. The MDX this way gets
a little complex. I have contemplated merging the date and time dimensions.
Does anyone have any guidance on maximum dimension size(in term of number of
rows)?

It depends on which version of AS you are using.

AS2000 - 64k limit
SSAS2005 - no limit

"Mukesh Kumar" <MukeshKumar (AT) discussions (DOT) microsoft.com> wrote

the number of maximum members in AS2K is limited to 64K for each child
members.
I mean 1 member can have 64K children in AS2K, like 1 city can have 64K
customers.
if you have more then 64K members, you can add a level between the city and
the customer level to split this 64K+ members into subsets of less then 64K
members (like the first letter of the name of the customer)

in AS2005 there is no issue like this.
does your 300 millions of rows are for a dimension or for a fact table?
if its a dimension, yes, its a big one...
from a fact table point of view, its a good starting point, but not so big.

from a date + time unique dimension, yes, its a good design.
you just have to make sure that your aggregations and partitions are
correctly setup to insure a good performance.



"Mukesh Kumar" <MukeshKumar (AT) discussions (DOT) microsoft.com> wrote

300 million rows are for the fact table.

I am using AS2005. With 2-3 years of detail date+time combined table will
have close to 2 million rows. Just wanted to make sure that it is not too big
for a dimension.

"Jeje" wrote: