If the Group By Style on the analysis parameters tab is set to User specified, a Group By element may be dragged onto the empty node under the Group By Clause, giving the following result. If User specified is not checked, a warning is given when the Group By Cube element is applied that the option will be automatically checked.
Columns can be moved into the folder titled Expressions (but not expressions or column positions), leading to a comma separated list of the columns that group rows in the report across multiple dimensions.
In the following example with the tutorial data table twm_customer, a group by cube of customer_count dimensioned and ordered by gender and marital_ status is requested.
The resulting data is listed in the following table.
| gender | marital_status | customer_count |
|---|---|---|
| 747 | ||
| 1 | 276 | |
| 2 | 353 | |
| 3 | 48 | |
| 4 | 70 | |
| F | 418 | |
| F | 1 | 159 |
| F | 2 | 189 |
| F | 3 | 30 |
| F | 4 | 40 |
| M | 329 | |
| M | 1 | 117 |
| M | 2 | 164 |
| M | 3 | 18 |
| M | 4 | 30 |
The blank spaces in the answer table above represent null values that signify the empty set rather than information that is missing. Through this technique, aggregations at many different levels are shown in a single answer table.
There are no special properties for the Group By Cube element.