have a fact table of 85000 records and about 5 dimensions. Only one
with 800 members a customer dimension. When I process the cube it first

reads the 85000 records from the fact table and then the number READ
rows keeps increasing to like 18 or 20 million records as having read.


I do not have a clue where Analysis services is finding this many
alarming number of rows. Appreciate how I can eliminate this kind of
weird reads.


Thanks
Karen

are you using AS2000?

this behavior occurs when you have a join between your fact table and a
dimension with a different granularity.
for example, you have a time dimension from year to day, your fact table is
only at the year level.
when you process your cube, a join is made between the time table and the
fact table and the result contain 365 * number of rows in the fact table.

try to optimize your cube (in the menu, you'll found this option)
disable levels in your dimensions which are not relevant for your fact
table.
verify your links between your tables.

"KarenM" <karenmiddleol (AT) yahoo (DOT) com> wrote

It could also be that a join is missing from the cube design or is just
joined to the wrong column. It is also possible that there may be an
issue with the data in one or more of your dimension tables.

If the join issue is not immediately apparent, one way to debug this is
as follows:

1) Get the "SELECT..." statement that AS issues against the relational
source either from the processing log file or from the processing status
screen in Analysis Manager.

2) copy the statement into query analyzer

3) comment out all the columns from the select and replace them with a
COUNT(*)

4) comment all the dimension tables and all of the join conditions in
the WHERE clause

5) run your statement - you should get the 85000 figure

6) add back dimension tables and the join conditions one at a time and
re-run the query.

7) when you hit the table or join condition with the problem you will
see the increased row count in the result of the query.

8) It is a good idea to check all the dimension tables as sometimes more
than one dimension can have issues.

HTH

--
Regards
Darren Gosbell [MCSD]
Blog: http://www.geekswithblogs.net/darrengosbell

In article <#AW1drR6FHA.3232 (AT) TK2MSFTNGP15 (DOT) phx.gbl>,
willgart_A_ (AT) hotmail_A_ (DOT) com says...