When documents contain multiple datasets, Intelligence Server joins the data in the datasets using a compound join.
If the datasets have any of the same attributes, the common attribute elements are matched.
Then beginning with the first row of each dataset and continuing to the last, a row is created in a virtual dataset, that is, the complete set of joined rows held in memory, consisting of all attributes, consolidations, custom groups, and metrics.
Metrics come from their respective datasets and exist only at the level of that dataset. The compound join saves memory space and processing time on the Intelligence Server executing the document.
When there are multiple datasets in a document, the grouping and sorting dataset is the one that is bold in the Dataset Objects panel. A document can be grouped and sorted using fields from the primary dataset only.
Related topics