Hi ​@himasha.abeywickrama,

The tabular model is deployed to the Analysis Service and during an upgrade nothing will happen to the deployed tabular model there.

The tabular model metadata, dataload/process definitions, entity criteria are data stored in the database and those should not affect during the upgrade process.

The scheduled jobs related to Analysis Models functionality is similar to any other scheduled job in IFS. Those will also be treated same during the upgrade.

Best regards,

Subash

Hi ​@subashfestus,

Thank you for the clarification!

One of our customers has a production environment setup for tabular models. However they are planning to replace the database with their IEE database during the go live. What will happen to the already deployed tabular models, environment setup and scheduled jobs in IFS? Should they be re-deployed/recreated since its a new database? Is there a way to retain those when deploying the new database?

Hi ​@himasha.abeywickrama,

Are there any tabular model configurations and data source configurations? If yes you have to export the configured objects before the DB replacement and then import again.

Data load/Process definitions will have to be recreated.
Environment setup needs to run again.
I’m not sure whether there is a way to export scheduled jobs. These tabular related scheduled jobs has to be treated as any other scheduled job.
Since the tabular model is already there in the Analysis Service there is no need to deploy them again as long as the SQL server is the same.

Best regards,

Subash
 

Thank you very much! ​@subashfestus 

​@subashfestus ​@himasha.abeywickrama  Our customer is going live and wants to ensure that the data in SQL Server is overwritten with the new data, removing the old test data. I’ve specified the process type as FULL in the process definition. Is there a way to ensure that this FULL process will completely replace the old data in SQL Server once they go live?

Hi ​@EntUviniG,

When you run the “Setup Tabular Model” assistant for a particular model, during the “Deploy Tables and views” step all the tables and views related to that tabular model will be dropped and recreated. When going live I would recommend doing this to get everything from a fresh state.

Also even without running that if you just want to load fresh data you have to use the “Full” load mode in the “data load definition”. “Full” data load mode will always TRUNCATE the table before loading data. Anyway for all dimensions always it will be a “Full” load. Only few facts support incremental loading.

After the data load you have to do a “Full” model process to get the new data.