In Pyramid, there are several different types of hierarchy, depending on the datasource and/ or the way in which the hierarchies were constructed in Pyramid. Each hierarchy type is associate with specific features and functions - driving the way their act in the tools and the capabilities they expose.
There are 2 main hierarchy type groups that are NOT mutually exclusive: structure and data. (For instance its possible that a hierarchy that is structurally a drill-path hierarchy is also a date-time hierarchy)
Although these terms are used interchangeability, usually 'dimension' refers to a collection of hierarchies - synonymous with a table in a database. Attributes refer to a single, flat hierarchy - synonymous with a column in a table. While hierarchies, in their purest sense are a collection of multiple levels that represent a tree like structure of elements - so its a collection of attributes.
Hierarchy Types by Structure
Structure definitions describe how the hierarchy is built and what it can do.
- Click here to see how to use Hierarchies when building queries.
Attribute hierarchies are "flat" hierarchies existing in the original data source, and containing member elements; each attribute hierarchy corresponds to a single column imported from the data source. The member elements contained within the attribute hierarchy correspond to the objects within the column imported from the datasource.
Regular hierarchies contain multiple attribute hierarchies, organized into levels, thereby enabling drill functionality. They also display expand and collapse buttons in grid visualizations, and display the drill path in the Elements panel. Regular hierarchies are usually even leveled - meaning every level in the hierarchy has the same element structure. However, these hierarchies can also be "ragged" - which means the levels are not even.
Drill Path Hierarchies
While drill path hierarchies are structurally the same as regular hierarchies, unlike regular hierarchies, they do not display expand and collapse icons in grid visualizations. Instead, the expand and collapse functions are accessed via the context menu, and the expanded levels are displays in a new column.
Drill path hierarchies also reflect the current hierarchy level in the drop zones and Elements panel, rather than displaying the drill path.
Parent-child hierarchies look very similar to regular hierarchies and typically include very similar functionality (most users wouldn't know or notice the difference). Parent-child hierarchies are more distinctive in HOW they are built, rather than how they are used. However, they can include some specialized aggregation logic and capabilities. ParentChildHierarchies defined in MS OLAP, SAP BW and SAP HANA datasources are also fully reflected in Pyramid; these hierarchies can be easily navigated, and can be used to create custom lists and formulas.
Unlike regular hierarchies, parent-child hierarchies are usually UNEVEN and ragged almost by design.
Hierarchies in both SAP BW and Microsoft Multi-Dimensional OLAP cubes can have hidden properties. The properties are, in effect, flat attributes that show different data values for the given member elements in the visible hierarchy. These items can be used in queries and visualizations, with some limitations.
Hierarchy Types by Data
Data definitions describe the data the hierarchy exposes and how it can be used in different calculations, logic or visualization.
GeospatialHierarchies are hierarchies containing geographical locations (like latitude, longitude, countries, states, cities, or zip codes) that have been assigned to a geospatial type. Geospatial hierarchies can be used to create map visualizations, in order to display your data set across a geographic location.
DateTimeHierarchies are hierarchies containing date-time member elements and have been assigned as a date-time type. They can be added to the query, or used to create classic date-time groupings (like year, month, week, etc), or to construct complex date-time calculations that can be saved and added to the query. They can also be used to create calendar slicers, where you can filter the query by date-time groupings or formulations that don't exist in the date model.
Hierarchies consisting of URLs can be added to the query, and the corresponding links (or the URL address) are displayed when the given visualization type is a grid or Cartesian chart. Click here to learn more.