There was a major change in delivering BI projects to Visual Studio 2019 and later. These projects are delivered as standalone Visual Studio extension and the SSDT installation from Visual Studio installer contains only SQL Server project type and some related tools.

It is still correct to install Visual Studion2019 with SSTD selected like in this picture:

But after that is done you should also complete installation of BI stack project types:

Download Reporting Services project type

Download Analysis Services project type

Download Integration Services project type

After installation will finish you can start Visual Studio 2019 and see that all your missing project types are there like in VS 2017:

Another important thing to remember is that with this different kind of installation is coming different update process: No more searching for SSDT updates on the web. You will receive notifications about updates for MS BI projects directly in Visual Studio:

This can cause a few issues when using source control like Azure DevOps across the team. If one the team members has enabled automatic updates installation in Visual Studio it can easily happen, that he will upgrade extensions which means that BI projects version will also be upgraded. After he will commit and other team members without updates to be installed will get the latest their project will be marked as Invalid (unsupported) and they must do an update to fix that issue. In a really large team it’s better to switch automatic updates off and manage this process centrally.

Download Azure Data Studio for all platforms.

After the quick installation process, there is a welcome screen with the option to download Czech localization, evaluate features in development, or opt-out from Microsoft data collection for product improvements.

The basic layout of the program looks like as in the picture:

There is Object Browser similar to Management Studio on the left side. The right side is dedicated to querying and results. Desing point of view looks like an attempt to merge UI elements from various platforms.

When a new connection is set up we can open the New Query window with IntelliSense, syntax highlighting, and the results section. All that very similar to SSMS. Cool improvement is the Peek Definition menu item which allows us to generate object definition and it works, surprisingly in contrast to SSMS, also for system objects like the view sys.tables.

Even if the Intellisense looks to be ok and similar to SSMS, it still won’t beat custom plugins like SQL Prompt from Redgate or SQL Complete from Devart for years.

We can also review the query execution plan. But it’s using a completely new set of icons and who is watching execution plans every day in SSMS will need some time to get familiar with the new layout.

The Results window is the same as in SSMS. Some core features like columns sorting and grouping are still missing.

The most interesting feature is hidden under the Charts tab. We have the option to visualize query results in few basic chart types directly. It can be extended by pressing the Create Insight button and store the chart definition for later use. What I don’t like so much is that we must edit the JSON template for this purpose. More details on  Insights.

Table data editor doesn’t bring much new compared to SSMS:

An interesting idea is the full integration of the terminal window which allows us to run various script as Powershell directly from Azure Data Studio.

Extensions module si similar to Visual Studio Marketplace and allows us to add various plugins and extensions with new features built by Microsoft or 3rd party developers. I see the SQL Server Profiler extension as the most useful one for database developer or DBA because it supports profiling Azure SQL Databases too.

If you don’t like something from the above mentioned and you would like to report it to Microsoft, then don’t hesitate to tweet it directly from the Studio. You can also report bugs easily or ask for missing features.

Step one, open the new SQL Server project and create one database called MainDatabase. In this database, create a view with the name [dbo].[ViewWithOuterDbReference] containing only one statement: SELECT * FROM [RererencedDatabase].[dbo].[Referencedtable]. This database doesn’t exist yet and that’s the reason why this object identifier is marked with a red line. Unresolved references are marked in the same way if we will import the existing database scheme.

Try building the project now. The result will be as seen in this picture:

Build engine did his job and discovered an unresolved reference and building failed because this reference can later cause later our queries to fail when referencing this view.

In our sample project, we need to take the following steps to build the project successfully.

Create another database in the same solution with the name [ReferencedDatabase]. Inside this database, create a new table called [ReferencedTable]. Now, we have something we can reference as [RererencedDatabase].[dbo].[Referencedtable] in queries from other databases.

However, if we will try building the project now, the same error will be fired because of our view [dbo].[ViewWithOuterDbReference] in [MainDatabase] still doesn’t know that the new database is there and the reference error cannot be solved. To move forward we need to explicitly set up a new database reference between our database.

Go to the References folder of [MainDatabase], right-click and select Add Database Reference:

The following dialog will pop up and we select the Database Reference as per the green box below:

Please note the Database Location section and Database variable field value. Go back to it.

After pressing OK, we can see a new item in the References folder saying, that our [MainDatabase] now has a new reference to [ReferencedDatabase]:

However, we are still not finished building will fail again. We have set the reference but next, we should adjust our code so that the build engine can use this new reference to solve existing reference conflicts.

By right-clicking on the new reference created above we will display its properties and copy the Database Variable Name property value to the clipboard:

Then, go back to the view and replace the [ReferencedDatabase] identifier with the [$(ReferencedDatabase)] variable from the clipboard. Shortly after that, a red line warning is removed from the code and we can build the project. All unresolved references have now been solved.

Further reading:

• SQL Server Data Tools (SSDT) and Database References