SSDT Import from Server deploy creates a duplicate AAS model

Question

I imported an existing Azure Analysis Services (Tabular) model into Visual Studio SSDT 2017 using Import From Server.

The project server and model properties point to the correct AAS instance.

After deploying, AAS shows 2 models (databases) : the original and a new one with the same content but with extra characters appended to the name. I only want to deploy changes to the existing model, not create a duplicate.

Why is a second model created on deploy?

Which SSDT/AAS settings control whether the existing model is updated vs a new one created?

How can I safely deploy updates to the original model and avoid duplicates?

Answer 1

Hello !

Thank you for posting on Microsoft Learn.

You pointed the workspace server of your SSDT project to the same AAS server you deploy to.

Tabular projects keep a workspace database for editing; SSDT creates it with a suffix (random/GUID-like characters).

When you then Deploy, SSDT also publishes a cleanly named deployment database. Result: two databases on the same AAS the workspace copy (with suffix) and the actual deployed model.

A second (less common) cause is that the deployment database name in project properties doesn’t exactly match the existing model name SSDT then creates a new database.

Identify the copies in SSMS, the one with the suffix is the workspace database.

The clean name is (likely) your intended/production model.

In SSMS, right click the suffix database then delete. (This does not affect your project files.)

Then, in Visual Studio: Project then Properties then Workspace Server then set to a local SSAS Tabular (developer/enterprise) or a dev AAS instance, not prod.

Then make sure that the deployment database matches the target and deploy.