Data Flashcards
Who defines measures and dimensions?
For relational - Tableau determines this automatically. By default, all numeric fields (that are not database keys) are measures, everything else is a dimension. The names of the fields are simply the names of the columns in the database. To change their placement just drag and drop.
For cubes - These are explicitly defined by the author of the cube.
Can I convert a measure to a dimension? Can I convert a dimension to a measure?
Yes. Drag a measure to the dimension pane.
For dimensions, either drag it to the measures pane if you are trying to count the dimension, or change its data type to number, if it’s the wrong type (right click > choose data type), and drag it to the measures pane.
What is a cube? Does Tableau support Cubes?
It’s a ‘multi-dimensional’ way to store data. The main benefit over relational databases used to be performance -typically a value is already calculated and stored for every combination of dimensions. Another benefit is increased structure such as hierarchies and predefined relationships.
Tableau can connect to cubes generated by Microsoft Analysis Services, Oracle Essbase (Hyperion), SAP BW
Note: Be prepared and able to demo against a cube/olap data source.
What’s the difference between using the generic ODBC connection to connect to my data, and the connections listed in the ‘connect to data’ dialog window? Are there drawbacks/benefits?
The connections listed in the ‘connect to data’ dialog window are optimised and tuned by Tableau to assure performance and full functionality of the full Tableau feature set. ODBC connections will rely on the driver (provided by database vendor) and may have limited support for the SQL standard. The result is that performance and functionality in Tableau may be degraded.
How do I join together sheets from two excel workbooks?
Tableau can only join sheets that are in a single workbook. You can use:
Edit > move or copy sheet
command in Excel to get all the sheets into a single workbook.
Data Blending will also work, but in many cases this will provide limited capabilities. If possible join.
Note: This is a trick question
Is Data Blending within a single database the same as using Multiple Tables in the connection dialog?
No, joins (in the multiple tables’ scenario) happen in the database with the raw data as part of one universal query. Blends happen in Tableau with the already summarised (aggregated) data. This means that Data Blending sends a separate query to each database. These queries aggregate the data. This means that Data Blending sends a separate query to each database. These queries aggregate the data to the same level of detail - regardless of the level of detail of the raw data. Then the results of the separate queries are blended together by Tableau.
Does Tableau always import the data?
No, and it is a differentiator with competitors. You have the option to use Tableau’s fast data engine or maintain a live connection to your database. If your data is already fast, or changes often, then direct connection is most likely best.
Is any functionality in Tableau dependent on having my data in the data engine?
No. Blending, calculations, table calculations, parameters, actions and all analytics are the same. However, some data sources will get extra capabilities when imported into the data engine (excel: count distinct and median)
What is a packaged workbook?
Packaged workbooks collect all the files related to the workbook and compress them into a single file. This package can then be delivered to other Tableau users without having to separately include data, images and other files referenced.
Does packaging automatically save all the data?
No. If your data is on a server, Tableau will not automatically extract the data. The reason is that the database could be impractical to pull down to the local machine. You can extract each data connection separately.
How do I schedule an extract to be refreshed?
Extracts can only be scheduled when using the Tableau Server. During the publish process, choose ‘Schedules & Passwords’ on the publish dialog. Choose a schedule and then publish.
