Analysis manager problem with huge tables

[angieh]

Hi All ,


am working with the analysis manager and it works fine unless the fact table contains huge amount of data .

is there any restrictions to the number of rows of the fact table ???

Thanks ,
Angie .

 

[MDXer]

I am aware of no limit to the number of rows the fact table may contain. What is the issue you are encountering? Paul

 

[angieh]

Hi Paul

First : Thanks for replying.

The problem is that when i try to add a fact table that have huge number of rows i got this message :
"cant count fact table rows please enter an approximate number of rows"

besides that when i try to make a dimension i got this error message :

"unable to open the recordset timeout expires"

i really dont know why and it soemtimes work fine with tables of small amount of data ..

Thanks ,
Angie

 

[pabloj]

I think you have to change the timeout parameter in the connection or on the database (max query time or something like that) also some index/database tuning could help.
Just my 2c Stick to your guns

 

[MDXer]

Definately index your fact but be sure the indexes are ones that will be used by AS, also make sure that all the STATS on the Fact Tale(s) are updated. To do this Build a partition and then copy the select Statement that AS issues to the SQL paste this select Statement into Query Analyzer and look at the execution plan. Other things to Check are make sure the cube is fully optimized. If a cube isn't optimized the dimension(s) that aren't or can't optimize join back to their dim tables. Also in the cube editor you can go to the Advanced properties of the cube and Manually put in the Approx size of you fact Table.

Are you using a star or snowflafe schema or are you attempting to build your dimensions directly from the fact Table? Paul

 

[angieh]

Hi paul,

first of all am trying to build dimensions from the fact tabel directly

second how can i view the sql statements issued by the AS ?

and please can u tell me links to find helpful resources for these issues ?

Thanks ,
Angie