So I've created a 2017 SSAS tabular model in VS. I've then deployed it to the SSAS tab work space server. But for some reason it creates two separate models. One normal and one with some random characters appended to it. Worse still, when I deploy changes to the model, it only updates the one with the characters appended.
What is going on here?
Deployment setting:
Two models showing in the work space server:
The tabular model with your username and a GUID is the workspace database. This is a local copy of the tabular model with the changes that you've applied to it when Integrated Workspace mode is not used. The workspace database is kept in memory while it's open in SSDT, and depending on the Workspace Retention property may be either removed from memory, removed from memory/disk, or still kept in memory. The default setting is to remove it from memory but not from disk, hence why you may not see this database whenever you close the model in SSDT. This property can by accessed from SSDT by highlighting the .bim file and viewing the properties (press F4).
The changes made to the workspace database should be applied to the deployed model when it's deployed using the model name as the target database on the deployment server, as in your screenshot. When you examine the model (non-workspace database) in SSMS, how do you know the changes are not applied to this, and have you tried refreshing the view from SSMS? This can be done by right-clocking the Database folder above the tabular models and pressing Refresh. Also, the deployment SSDT is succeeding without errors, correct?
Since I cannot comment on the correct answer I'll make it a separate answer.
What I saw happening was when you import a cube in Visual Studio (while creating a new project) a new tabular DB is created with the name "CubeName_username_Guid".
Everything you modify gets deployed to the real cube after you close Visual Studio. Also after you close VS, the oddly named cube also disappears.
I hope it helps.
