Catalog Vocabulary

The Fabric Catalog introduces a vocabulary that describes the catalog entities and the relations between them. The relations indicate the connections between the data entities and determine their hierarchy.

The below vocabulary serves as a model for describing a Catalog and assists with processes standardization across different interface types.

The data entities are represented by nodes and the referential links between the nodes are represented by relations. Nodes and relations have predefined properties that enrich the Catalog schema.

Additionally, due to differences between the data sources, some nodes' properties are generic, while others are relevant only for specific interface types.

The following 2 tables describe the node and relation types, and how they are defined in each interface type:

Node Types

Node Type	Fabric Interface Type
dataPlatform	Represents a Fabric interface in the Catalog data model
schema	Represents a schema of the Fabric interface
dataset	Represents a dataset (e.g. table) of the Fabric interface’s schema
class	Represents a dataset (e.g. table) of the Fabric interface’s schema (1:1 to dataset)
field	Represents a dataset field
field	The field data type can be: primitive - string, integer, blob, date, number, boolean or any. collection - an array of primitive values.

Relation Types

Relation Type

Description

contains

dataPlatform contains schema

schema contains dataset

dataset contains class

class contains field

The direction is One To Many.

refersTo

dataset refersTo dataset (PK field)

Example: INVOICE refersTo CUSTOMER (customer_id)

The direction is Many To One. The PK / FK columns are included in the relation's properties.