Introduction
What are Shared Data Sources?
Data Sources are one of the two objects that can be created in Tableau Desktop and shared on the Tableau Server. They are basically an XML file with information on your data source including connection information, tables, joins, and calculations. They can also have embedded credentials and extracts included.
Why should I use Shared Data Sources?
They make life easier. One of the underutilized (at MM at least) features of Tableau Server; many people don't know about them or how to use them.
Some Helpful Links
Instructions
Use Case One - Starting Points
When you create a data connection in Tableau desktop it has all the fields, columns, and calculations you want. You can save a .tds file and reuse it so you don't have to go through the hassle of recreating it. You can also publish it to the server for others to use.
If there is a change in the table structure or other changes to be made pushing them out to others is easy. You can make any changes and republish the data source. The next time someone opens a workbook using it it will notify them there is an update and ask them if they want to use it.
Permissions are usually set to allow users to download, but not connect. In this case no extract or credentials are included. The big benefit is reducing repetition and errors.
Use Case Two - Speed up development
If you include an extract and schedule it to refresh you can speed development two ways.
- If the extract is large it is separated from the workbook. Changes to the workbook can be republished without having to push all the data across every time.
- The data in your workbook is always updated. If you have an extract in the workbook it needs to be regenerated (or the workbook re-downloaded from the server) every to have the most recent data. Having a subscribed data source means the server handles the updates for you.
Use Case Three - One extract, many workbooks
Two extracts with the same data takes twice the space and resources to extract. Having one data source and multiple workbooks using it means a lot less server space and load. Oh, also less time developing the workbooks.
Use Case Four - Self Service BI
Did you know that you can publish a DataSource and let people make new workbooks using it on the web? No additional license costs, and it has about 80% of the features of Tableau Desktop.
What to watch out for
Best practice is to create and edit the data source in a empty workbook to prevent dashboard specific components from bleeding over.
Calculations can be part of the data source or part of the workbook. If a calculation is going to be useful to others you can include it in the data source. Any Calculations that use Parameters should be created in the workbook as the parameter isn't part of the datasource.
If using an extract some resource intensive calculations should be included in the data source since they will be calculated at the time of extract instead of on the fly. The exception to this is calculations that use parameters or functions like today().