A semantic mapping or linking is a relationship between two objects in the
It consists of one or more mapping elements, which are
relationships defined between a destination element and one or more source elements.
general, a semantic mapping describes how elements in a source model (more conceptual) define
elements in a destination model (closer to an implementation or representation). Elements in
the destination model are representations or implementations of the associated element in the
You can perform two types of semantic lineage:
- Semantic usage: From the more general or design to the more specific or implemented
- Semantic definition: From an implementation or specific metadata element to its design
or defining term (up).
You can also perform two types of linking:
|Type of semantic linking
||Links from a term to another metadata element.
You define a business name by
creating a term in a glossary, such as Email address. You reuse
the term to apply the business name to related elements, such as to any columns or
fields with email addresses.
Semantic links can be used for the following use cases:
- More general to more specific
- Design to implementation
- Defining term to defined element
Classification is the primary and recommended method as it is system-managed.
It takes precedence over semantic mapping for lineage (definition lookup and usage).
Classification works only with glossaries at the more abstract end of the link, using the
Classify feature from the Semantic Flow tab.
As it is managed by the system, you do not need to create mapping contents between the
glossary and the target model. You do not need to manage the semantic lineage lines in the
For more information on classification, see Classifying a metadata element with a term.
Semantic mapping method
Semantic mapping is the secondary method as it is more user-managed.
Semantic mapping can be defined between any two contents such as glossary and/or
You must explicitly create and manage a content for each mapping content between two
models/glossaries. Once created, you can map to individual data elements as much as you want
within the scope of those two models.
As there is existing content in the repository, it can be managed using the semantic
mapper, be exported and re-imported using the CSV format or be "embedded" to be migrated
from semantic mapping to classification from a glossary.