Creating Hierarchies

Use the Hierarchies page in the Define Sources tab of the Admin module to create hierarchies and assign level keys to a hierarchy level.

A Hierarchy defines the levels of analysis and aggregation along a dimension. A hierarchy defines multiple levels that describe relationships between attributes, and the attributes that belong to each level. For example, in a Location dimension a hierarchy could be defined as Region - Country - Province - City. A hierarchy allows report users to click from higher levels into lower levels of detail.

A level key is one or more columns of a single dimension in a dimensional hierarchy that uniquely identifies records at that dimension, at that level of the hierarchy. Also called a "level key column".

Notes About Creating Hierarchies

You can only create hierarchies for Advanced spaces. See Space Types.

Birst auto-generates the Level Cardinality value. You can assign a different value if desired. See Navigation and Cardinality.

The Advanced button is only enabled for users who are members of a group that has the AdvancedAdmin ACL. Use this button to access the Advanced Properties dialog box in order to create a hierarchy alias or set a hierarchy as a Type 2 Slowly Changing Dimension.

For more information about hierarchies, see Carrying Higher Keys Downhill, About Member Attributes and Floating Levels and Case Study: Creating Grains and Hierarchies.

Level Names in Multiple Hierarchies

Birst naming conventions no longer enforce having unique level names within two different hierarchies, with the following exceptions:

If the hierarchy is H1->L1 and H2->L2 and data is already processed once and then level name is changed to H2->L1, the new table would be generated in this case but data from the old fact/dimension table would not be carried over.

If the hierarchy is H1->L1 and H2->L1 and data is already processed and then hierarchy changed to H2->L2, then when the repository is rebuilt, a new fact table would be introduced and again, data from the old fact/dimension table would not be carried over.

Birst recommends using non-unique names only on new spaces and only in the case of a specific business requirement that enforces such a level naming convention.

To create new hierarchies and levels within a hierarchy

Go to Admin - Define Sources - Hierarchies.

Tip: A hierarchy and its levels are not linked to a specific data source on the Hierarchies page (see image below). Data sources are displayed only to make it easier to find columns that represent the hierarchy level keys. After hierarchies and levels have been set up, if you select a hierarchy or level on the left, the data source that is highlighted in the data source box may not be the original data source that was used to select the level keys, however, the selected source will contain the level key columns that were assigned. The level key columns can be changed as necessary without regard for the data source that is used.

Click the down arrow on the Add button and select Add Hierarchy. A hierarchy called New Hierarchy will appear at the bottom of the Hierarchies list. If you expand New Hierarchy, you will see New Level.

Tip: To change the name of a hierarchy or level, click the hierarchy or level and type a new name when the text is highlighted.

To add a level within a hierarchy, select the hierarchy then click the down arrow on the Add button and select Add Parent Level or Add Child Level.

To assign level keys, select the hierarchy level from the left pane.

In the upper right pane use the search box to locate a source that contains the level keys. Highlight the data source.

In the Level Key Columns pane in the lower right, check each column that represents the level key for that hierarchy level.

Click Save.

Tip: To revert back to the latest saved changes, click Cancel.

Next Steps

After the hierarchies have been created, you can assign hierarchies to data sources using the Grain tab of the Manage Sources page.

See Also
Creating a Data Model
Creating a Hierarchy Alias
Setting a Hierarchy as a Type 2 Slowly Changing Dimension
Creating a Custom Time Hierarchy
Defining the Grain of a Data Source
Carrying Higher Keys Downhill
About Member Attributes and Floating Levels
Case Study: Creating Grains and Hierarchies