Understanding subsets

Conceptual information about subsets.

When using a subset, please keep the following things in mind:

• All of your benchmarks now relate to your subset, not to the entire dataset, which is much more useful when analyzing a specific subset. See Understanding Benchmarks.

• Using a subset affects all of your workspaces because the subset is applied globally to Data Workbench.

• Subsets affect only metrics and denormal dimensions, not normal dimensions.

• When using Report, subsets do not affect the data in reports published for others to view.

• Once applied, your subset is in effect for all subsequent work in the profile, including the next time that you open this instance of Data Workbench, until you remove it.

• The only place that indicates that a subset has been applied is the context menu that you reach by right-clicking within a workspace.

  

• You must be working online to change or remove a subset. If you are working offline and have applied a subset, you cannot view results from the entire dataset. See Working Offline and Online.

  NOTE

  The size of your subset is limited to the amount of data in your filter that resides on a single Data Workbench server. Therefore, if your dataset spans a Data Workbench server cluster, data for your subset comes from only one Data Workbench server in the cluster.

A user at a large retailer wants to create a subset (local cache) of one particular work week of data and then run queries only on that week of data. To do this, the user creates a subset for the days of interest.

The following example shows you a bar graph of Visitors over time and a Traffic metric legend. The first figure contains no selections: all of the data in the dataset is represented. The second figure shows data for a subset of Days = {…} by Visitors, in which Days is based on a selection of the elements April 1st through April 5th in the Day dimension.