Master Data Services Overview
Applies To: SQL Server 2016 Preview
In Master Data Services, the model is the highest level container in the structure of your master data. You create a model to manage groups of similar data, for example to manage online product data. A model contains one or more entities, and entities contain members that are the data records.
For example, your online product model may contain entities such as product, color, and style. The color entity may contain members for the colors red, silver, and black.
Models also contain attributes that are defined within entities. An attribute contains values that help describe the entity members. There are free-form attributes and domain-based attributes. A domain-based attribute contains values that are populated by members from an entity and can be used as attribute values for other entities.
For example, the product entity might have free-form attributes for cost and weight. And, there is a domain-based attribute for color that contains values that are populated by the color entity members. This master list of colors is used as attribute values for the Product entity.
Derived hierarchies come from the relationships between entities in a model. These are domain-based attribute relationships. In the product model for example, you can have a color derived hierarchy that comes from the relationship between the color and product entities.
Once you've defined a basic structure for your data, you can start adding data records (members) by using the import feature. You load data into staging tables, validate the data using business rules, and load the data into MDS tables. You can also use business rules to set attribute values.
The following table outlines the key Master Data Services tasks. Unless otherwise noted, all of the following procedures require you to be a model administrator. For more information, see Administrators (Master Data Services).
You might want to complete the following tasks in a test environment and use the sample data provided when you install Master Data Services. For more information, see Deploying Models (Master Data Services).
Create a model
When you create a model, it is considered VERSION_1.
Create as many entities as you need to contain your members.
Create entities to use as domain-based attributes
To create a domain-based attribute, first create the entity to populate the attribute value list.
Create attributes for your entities
Create attributes to describe members. A Name and Code attribute are automatically included in each entity and cannot be removed. You might want to create other free-form attributes to contain text, dates, numbers, or files.
Create attribute groups
If you have more than four or five attributes for an entity, you might want to create attribute groups. These groups are the tabs that are displayed above the grid in Explorer and they help ease navigation by grouping attributes together on individual tabs.
Import members for your supporting entities
Import the data for your supporting entities by using the staging process.
For the Product model, this might mean importing colors or sizes.
You can also create members manually.
Create business rules to ensure data quality
Create and publish business rules to ensure the accuracy of your data. You can use business rules to:
Import members for your primary entities and apply business rules
Import the members for your primary entities by using the staging process. When done, validate the version, which applies business rules to all members in the model version.
You can then work to correct any business rule validation issues.
Create derived hierarchies
Derived hierarchies can be updated as your business needs change and ensure that all members are accounted for at the appropriate level.
If needed, create explicit hierarchies
If you want to create hierarchies that are not level-based and that include members from a single entity, you can create explicit hierarchies.
If needed, create collections
If you want to view different groupings of members for reporting or analysis and do not need a complete hierarchy, create a collection.
Create user-defined metadata
To describe your model objects, add user-defined metadata to your model. The metadata might include the owner of an object or the source the data comes from.
Lock a version of your model and assign a version flag
Lock a version of your model to prevent changes to the members, except by administrators. When the version's data has validated successfully against business rules, you can commit the version, which prevents changes to members by all users.
Create and assign a version flag to the model. Flags help users and subscribing systems identify which version of a model to use.
Create subscription views
For your subscribing systems to consume your master data, create subscription views, which create standard views in the Master Data Services database.
Configure user and group permissions
You cannot copy user and group permissions from a test to a production environment. However, you can use your test environment to determine the security you want to use eventually in production.
When ready, you can deploy your model, with or without its data, to your production environment. For more information, see Deploying Models (Master Data Services).