cancel
Showing results for 
Search instead for 
Did you mean: 

Adding new dimensions to an exiting model

former_member196865
Participant
0 Kudos

Hi Gurus,

I am trying to add new dimensions to an already existing model and in this process before adding the dimension, I tried to export the data from the model using the data manager package for Exporting. While doing this, the data manager is throwing an error that the the data intersections have exceeded the 1,000,000,000 mark. I have tried it for one month alone too and it still says the same. What could be the reasons for this? Can I export the data from the model using other means?

Regards,

David

Accepted Solutions (0)

Answers (1)

Answers (1)

former_member186498
Active Contributor
0 Kudos

Hi David,

instead of export package use SSMS to copy the fact table but before do a full optimize if you have data on fac2 and factwb.

Regards

     Roberto

former_member196865
Participant
0 Kudos

Hi Roberto,

Thanks for your reply. I am not sure what you are trying to say. Do you mean that I have to copy the fact table and then add my new dimensions to the existing model and then copy the data back to the old one or how does that work?

don't we have to add dimension header data when importing the data back?

Regards,

David

former_member186498
Active Contributor
0 Kudos

Hi David,

you can copy the table with a "insert into tblcopyfact select <all the fields> from tblfact<Model>"

add your dimension with at least one record inside, full process
and replace the data with a reverse "insert into" adding in the right place of the field list, the default value for the new dimension and full reprocess the dim and after the modify model with process application to rebuild the OLAP.

Regards

     Roberto

former_member196865
Participant
0 Kudos

Roberto Vidotti wrote:

Hi David,

you can copy the table with a "insert into tblcopyfact select <all the fields> from tblfact<Model>"

add your dimension with at least one record inside, full process
and replace the data with a reverse "insert into" adding in the right place of the field list, the default value for the new dimension and full reprocess the dim and after the modify model with process application to rebuild the OLAP.

Regards

     Roberto

Hi Roberto when you say "add your dimension with at least one record inside, full process" Where are we adding this dimension? In SSMS or BPC? I have created the dimensions in BPC but I haven't added them to the model yet. I have processed them in BPC.

So my understanding of you what you are saying....

1) Create a dimensions with atleast one record in BPC and process them in BPC

2) Create a copyFacttable so that data from facttable is copied into.

3) Use Insert Into tblcopyfact select * from tblfactmodelname

4) Add the dimensions into the model from BPC

5) Optimize the model in BPC

6) Add columns into the copyfacttable we created earlier for the inserted dimensions

7) Add a some records into the columns we created

😎 Send this data into the Facttable using insert into tblfactmodelname select * from tblcopyfact

9) Reprocess the dimension in SSMS

10) Optimize the model in BPC

Let me know if I am saying this right.

Regards,

David

former_member186498
Active Contributor
0 Kudos

Hi David,

before starting do a full optimize (compress data and free fac2 and factwb)

1) yes you should create the dimension in BPC and adding one record as default membervalue <DIM>_NOAT,

ok for the steps 2, 3 and 4

instead of 5) do a modify application selecting "create sql index" and "modify application", if you have inserted the default membervalue the modify application should recreate the model and mantain the fact table adding the default value to the new dimension: in this case your work is finished.

Otherwise, only if modify appl. deletes the fact table:

[continue with steps 6, 7 (normally add just default value) and 😎

instead of 9 and 10 (optimize is useless because you have done it before) do again a modify application with creation of SQL index and process application, this with recreate the OLAP.]

Regards

     Roberto