Add a Snowflake Dimension
In AtScale, a snowflake dimension refers to a logical dimension comprised of columns coming from more than one physical dataset.
note
Note: Unhandled NULL
values in key columns will result in incomplete
aggregate tables and unexpected query results. See the Custom Empty
Member
feature for more details.
Procedure
To add a snowflake dimension:
- Click the plus icon in the Dimensions panel to add a new empty dimension.
- Enter a name for the new dimension.
- Find the new dimension on the main canvas and double-click to open it in the dimension editor canvas.
- Drag all of the datasets that the snowflake dimension will use from the Library panel to the dimension canvas.
- Click the plus icon in the Hierarchies panel to add a new hierarchy.
- Drag the column that represents the dimension's key level attribute just below the hierarchy name in the Hierarchies panel. This creates a new level based on that dataset column.
- Continue adding levels by dragging dataset columns just above the previous level in the Hierarchies panel. A hierarchy should go from the most granular level on the bottom to the most general level on the top.
As you add a level from a column of a different dataset, AtScale will
automatically create the level relationship if the dataset in the level
below has a column of the same name. For example day.monthkey
rolls up
to month.monthkey
. If the relationship cannot be created, you'll have
to choose the dataset columns to link the new level to the one below it.