Thanks to everyone who attended my webinar on the ins and outs and the basics of building your first Tabular Model. I hope everyone learned a little something and maybe picked up a couple tips or tricks along the way.
Click Here To View The Recording
And thanks to everyone who had questions! I apologize for not being able to get to the questions during the webinar, so here are answers to a few of the questions posted in the chat window.
Question: “Don’t use joins” – that includes views that have joins under the covers, right?
Answer: That’s correct. That’s still going to create additional stress on the relational engine at query time and the last thing you want is to run into a locking/blocking issue during processing because you have complex queries behind the tables in your Tabular Model. If you need to join to other tables to pick up other columns, I would suggest getting with your ETL people and materializing the desired columns in the table so you can eliminate the joins.
Question: Could you please explain what difference of perspective and role?
Answer: A Role is used to define member permissions for your model. The Role defines what actions a user can perform on the model, such as read or process. A Perspective is a viewable subset of the model. A Perspective is similar to a SQL Server View on a Table. It can be used to display a subset of the model to a business unit in order to make it a little more simple to navigate through the model.
Question: What is the difference between Active & Inactive relationships?
Answer: In a Tabular Model, multiple relationships can exist between two table, but to create accurate calculations, we need a single path from one table to another. This means that even though multiple relationships exist between the two tables, only one will be actively used. The inactive relationships between the tables will not be used.
Question: Can a tabular “object” once built be put in source control and deployed by IT like SSIS packages and SSAS cubes?
Answer: Yes, Tabular Models can be deployed using the SSAS Deployment Wizard. The Create Script can also be generated from a deployed copy of the database and used to deploy to another server. For more information on SSAS Tabular deployment methods, see here.
Question: is there a way to change the database name and remove the ugly guid?
Answer: The guid is only appended to the database on the workspace server. The database can be deployed with whatever name you like. You can set the database name by going to the project properties and setting the Database property. Whatever you enter in as the Database name will be the name of the database once you deploy the database and will not include the yucky guid, unless you company standards are to use the guid :).
Questions: Does Tabular model have superior performance over OLAP based cubes? Also can Tabular model be used for time series data?
Answers: I would say generally speaking that a Tabular Model will have much faster query response on average, which is definitely one of the strengths of an in-memory database. OLAP cubes have the ability to contain much more data and can scale out better but in my experience queries against Tabular Models are usually much faster than queries against Multidimensional databases. Tabular Models also allow for time calculations. For more information on the kinds of time intelligence functions and calculations you can create with DAX, check this out.
Once again, thank you so much to everyone who attended my webinar. If I didn’t get to answer your question, I apologize. Feel free to post your question in the comments or on the forums of BIDN.com.