Showing posts with label olap. Show all posts
Showing posts with label olap. Show all posts

Monday, February 20, 2012

Problem with parent-child dimension

I have to migrate OLAP database from AS2000 to AS2005. I have a problem with one of my dimension. This is how the situation looks like.

In AS2000 I have a parent child 'Users' dimension. Here is the structure of this dimension

Users

UniqueIdUser - Key

UniqueIdUserPrev - Parent

IDUser

UserName

The parent-chld relation is based on colums : UniqueIdUser and UniqueIdUserPrev. In AS2000 connection between this dim and the facts table is based on dim.IDUser - fact.IDUser. I know that this kind of connection gives me cartesian since dim.IDUser is not unique, but this is what I expected.

The question is : How can I do a parent child dimension in AS2005, where a parent child relation will be based on UniqueIdUser and UniqueIdUserPrev colums but connection to the fact table will be based on IDUser?

How about using a named query, which includes UniqueIdUser, for the fact table, like:

select fact.*, dim.UniqueIdUser

from fact

join dim on fact.IDUser = dim.IDUser

Corrsepondingly, the dimension table could now be a named query, like:

select distinct UniqueIdUser, UniqueIdUserPrev, UserName

from dim

|||

Hi

Thx for advice. Yes It will work, cube can be based on view like this. But my fact tables are quite big - something about 5-8 millions as well as dimUsers (about 20K). Inner join between this two tables will not work efficiently. There must be another solution. I have migrated cubes from AS2000 to AS2005 and AS2005 managed to convert this cube. But I can not do it manualy. AS2005 has created wired relation in 'Dimension Usage' between UniqueIDUser from 'Dimension Colums' and UniqueIDUser from 'Measure Group Colums' although there is no column UniqueIDUser in 'Measure Group Colums' (You can not choose it from drop down list)!!? If wizard could do this, I should mange to do this manualy also.

|||

Maybe you can inspect the migrated .cube XML file, to find out how the Granularity attribute is configured (there should be an element like: <Type>Granularity</Type>).

My understanding is that not all of the XML elements generated by the Migration Wizard can be manually recreated via the BIDS GUI. Of course, you could edit the .cube XML file directly, but then you'd have to understand it very well.

problem with panda antivirus and olap cubes

I have the next problem:
i have MS sql server 2005 installed in a server.
We have some olap cubes in the server deployed with analysis services 2005.
When we try to connect to the analysis server from a client pc using pivot tables we obtain this error:
"Errors in network layers. Error code=0XC10D0009, External code=0x00000000"
The client pc has panda antivirus (corporate) with true prevent
When i unisnstall it, we can connect to the analysis server and to the cube normally and when we install the antivirus again (without true prevent and with the minimun options) the problem starts again and we can't conect to the analysis server.
Anybody had a similar problem?
Anybody knows a solution for it?
Thanks
Jose
coormx@.yahoo.com.mx

Jose,

In this situaion, the fast way is to contact the antivirus vendor to solve the issue.

However, if you can post your *connection string* and *exact error message*, somebody in this forum might help you to pinpoint your issue.

|||

Moving to the SQL Server Analysis Services Forum.

|||panda has fixed the problem. Tank you panda

problem with panda antivirus and olap cubes

I have the next problem:
i have MS sql server 2005 installed in a server.
We have some olap cubes in the server deployed with analysis services 2005.
When we try to connect to the analysis server from a client pc using pivot tables we obtain this error:
"Errors in network layers. Error code=0XC10D0009, External code=0x00000000"
The client pc has panda antivirus (corporate) with true prevent
When i unisnstall it, we can connect to the analysis server and to the cube normally and when we install the antivirus again (without true prevent and with the minimun options) the problem starts again and we can't conect to the analysis server.
Anybody had a similar problem?
Anybody knows a solution for it?
Thanks
Jose
coormx@.yahoo.com.mx

Jose,

In this situaion, the fast way is to contact the antivirus vendor to solve the issue.

However, if you can post your *connection string* and *exact error message*, somebody in this forum might help you to pinpoint your issue.

|||

Moving to the SQL Server Analysis Services Forum.

|||panda has fixed the problem. Tank you panda