Taxonomy enables us to perform a quick search and locate a few specific MDM records in a database of up to a million records.

The purpose of taxonomy is to classify like master data records together into categories, based on a set of common, category-specific characteristics.

In MDM, taxonomy is a hierarchical representation that allows categories to be part of other categories. As we progress down a taxonomy, the type of records included are narrowed down based upon the category.

MDM utilizes this hierarchical concept of taxonomies to represent them as a tree.

It supports eClass taxonomies, UNSPSC, and custom classifications.

As we progress down taxonomy, every category has a defining characteristic in addition to the ones corresponding to every category above it in the hierarchy. For instance, in the taxonomy of rocks, igneous rocks have certain specific attributes such as color, texture, and chemistry as well as common characteristics (such as weight and shape) shared with other types of rocks.

In MDM terminology, characteristics are known as attributes and represent fields of information applicable to a handful of main table records in the repository.

Every taxonomy table has a pool of attributes which can be linked to one or more individual categories. Using the Taxonomy mode, you can carry out the management of the attributes in the taxonomy pool.

Attributes linked to categories are associated with main table records only when a record is assigned to a particular category within the taxonomy. The attributes of the parent category are automatically inherited by the main table record assigned to the child category.

A main table record, therefore, consists of common fields, inherited attributes, and category-dependent attributes.

Qualified tables allow extreme versatility in MDM. These are special kind of lookup tables that can store complex relationships. These relationships exist between the main table record and one or more lookup table records that contain other additional information.

A qualified table contains fields called qualifiers whose values do not apply to the qualified table record. The value of a qualifier is specific to the association of each main table record with the qualified table record. Qualified tables support for multiple prices (based on regions or types including quantity price breaks), cross-reference part numbers, other distributor/supplier/customer-specific information, and product applications.

For example, consider a scenario where a product has multiple prices based on its region. We can utilize the qualified tables to support this scenario out-of-the-box.

In a situation when the contents of an MDM repository needs to be published containing product information, a finer organization of the records is required than that provided by the categories of a taxonomy. Such as grouping main table records based on not only the category but also the manufacturer. Product families enable support for such granular classifications of records such as eClass taxonomies or UNSPSC within an MDM repository.

A group of main table records can be related based on the value of one or more common fields and/or attributes. Such related main table records form a product family which contains additional family data fields such as image, logo, description, specifications, and so on.

A single master repository can be sliced into a number of custom virtual repositories with the help of product masks in MDM. This considerably simplifies the maintenance of a single repository which needs to be targeted to multiple audiences.

A user sees each virtual repository as a completely private repository containing a subset of the records from the master repository.

Virtual repositories created using product masks can, for example, be used for custom subsets for contact customers and targeted market segments driven by a single MDM repository.

The added advantage of product masks is that they impose no performance limitations on the DBMS as they are defined at the individual product level rather than at the query level (in contrast to SQL views that are defined at the query level).

In the electronic catalog's web page, a mask can automatically be applied upon site entry thereby controlling the view of the user to see only a part of the MDM repository that is relevant.

An MDM repository containing a large volume of images, text blocks, and PDF files needs to have an organization mechanism which enables them to be easily located and linked to a particular master data record. Data groups allows such organization wherein each object is assigned to a data group when it is first created or imported into the system.

Data groups can further be organized in a hierarchy similar to the taxonomy hierarchy. Thus data groups allow a parallel classification scheme like taxonomy hierarchy to organize objects into subgroups called data groups. Individual data groups can exist for product images, category icons, and manufacturer logos.

Validations in MDM are similar to formulas in Excel that return a Boolean result to denote success or failure. Validations include the following capabilities:

Validations are created and executed using the MDM Client. A user can run complex tests for different types of conditions against a group of one or more records.

Validations can be branched based on a specific condition in the main validation. MDM automatically executes the relevant validation based on the condition such as the category value of each record.

Validations can be aggregated and grouped into validation groups and executed as a batch. Thus, multiple related validations can be grouped under a single validation group and run together by running the validation group against the master data records. This simplifies the task of multiple validations one-by-one and reduces time and effort.