Managed attributes

Managed attributes are similar to standard attributes in that they describe the records in your data set.

Unlike standard attributes, (which only provide a way to assign values on records in your data set), managed attributes allow you to capture additional characteristics that may be present in your data. Once captured and loaded into a running Dgraph, these characteristics become part of that data domain.

Managed attributes allow you, as a data architect, to capture the following characteristics of your records:

A set of predefined allowed values. Some attributes on your data records may have a requirement to have assignments only from a predefined set of allowed values. For example, an attribute currency may have a predefined set of values (dollars and euros). A managed attribute allows you to define a set of specific values that are allowed on a standard attribute.
Hierarchy. Some attributes on your data records may benefit from being organized in a hierarchy. For example, an attribute representing a ProductCategory type of Clothing may have different types of clothing items underneath it, each represented by a standard attribute (such as Caps, Gloves, Jerseys, and so on). A managed attribute allows you to define the hierarchy of standard attributes.
Additional metadata on attribute values. Finally, your source data records may have attribute values that could include additional metadata, such as text descriptions. Managed attributes allow you to capture these additional metadata on attribute values.

Managed attributes are often used to support hierarchical navigation. In other words, associating a managed attribute with a standard attribute enables hierarchical navigation of records based on the standard attribute values. For example, you can navigate a collection of books using the Library of Congress Classification standard attribute, and refine by Literature > American > 19th century. (Note that while managed attributes can capture the hierarchy of your attributes, they are not required to contain hierarchy information.)

When you create a managed attribute whose purpose is to represent a hierarchy, you load a taxonomy definition that enumerates a hierarchy where each standard attribute value (in a key-value pair for the standard attribute) is a node in the hierarchy (called a managed attribute value, or mval).

Managed attributes are described by system records — PDRs and DDRs.