Hi,

I have a production server with a database deployed and the dev. server
based on a subset of data.
On the production server I do some design aggregation to improve the
performance, while this is not required on the dev. server.

But when I deploy a new version of my database I lost my aggregations.
most of the time the new version provides new calculated measures and
perspectives, the main cubes / dimension doesn't change.
So I want to keep the aggregations.

what can I do?
how to save my aggregation design of my production version?

thanks.

Jerome.

When you deploy a BI project (from Visual Studio), you are overwriting the
objects on the server and there is no setting to merge the Cube from the BI
project with the Cube from the server, keeping the server's aggregations.


One work-around is to re-create the BI project from the server (using the
"Import Analysis Services 9.0 Database" wizard from File -> New Project;
please double check the project name to match the database name). But we
moved one problem to another: this new BI project has aggregations that you
don't want on the dev server.


Another work-around would be to script the cube definition from the
production server (using SQL Management Studio, right click on the cube item
and use 'Script Create' or 'Script Alter'). Then create a new Batch script
in which you copy similar Create/Alter stamenents for all the
AggregationDesign objects. It will look something like this:

<Batch xmlns="http://schemas.microsoft.com/analysisservices/2003/engine">
<Create AllowOverwrite="true">
.... the AggregationDesign 1 ...
</Create>
<Create AllowOverwrite="true">
.... the AggregationDesign 2 ...
</Create>
...
<Create AllowOverwrite="true">
.... the AggregationDesign n ...
</Create>
</Batch>

Please script the Create/Alter statements for other objects to get sample
XMLs. When you have the complete <Batch> script, you can run it with AMO or
a DTS package on the production server after you deploy your BI project. In
short: deploying the BI project continues to wipe the aggregations, but then
you create them back with the script. Unfortunately, any data for those
aggregations is still lost (but if you did structural changes to your cube,
you would need to re-process it anyway after deploying the BI project).

Adrian.



"Jéjé" <willgart_A_ (AT) hotmail_A_ (DOT) com> wrote