How do different schema definition languages
Answers
Answered by
0
If you're referring to this Three schema approach:

most data-related description languages or tools that are concerned with schemas focus on the "physical level" and "view level", with the "conceptual level" mostly used in putting together the schema design itself.
The physical model in most relational databases is described using SQL
DDL. In NoSQL worlds, physical schemas can be attached to particular records using JSON or XML in the fashion of MongoDB or MarkLogic, or more generally described using a kind of DDL in something like Apache Cassandra.
There are languages like UML that try to describe conceptual schemas, but these aren't widely used these days, although they were a "hot thing" for a time in the mid and late 1990s. For the most part in the relational world, the closest thing to a "language" for a "conceptual schema" is probably Entity–relationship model diagramming and description tools. Tools like Embarcadero Technologies E/R Studio can reverse-engineer a ER diagram from an existing SQL schema and generate SQL DDL from a user-created ER diagram.
The "view level" or "external level" is typically implemented outside of data managers in user-side code, possibly using Object-relational mapping tools like Hibernateto convert structures that are in "physical layer" schemas to more "external-friendly" structures that ultimately map to display or user input UI elements in application code. Other code may just have a "database layer" that completely implements this interface inside the API (and a lot of code will use both a "database layer" API as well as OR mapper tools and libraries).
(Frankly, I don't think database views really are "view layer" things, as they are still quite "physical" in their implementation and behavior. But others may disagree...)

most data-related description languages or tools that are concerned with schemas focus on the "physical level" and "view level", with the "conceptual level" mostly used in putting together the schema design itself.
The physical model in most relational databases is described using SQL
DDL. In NoSQL worlds, physical schemas can be attached to particular records using JSON or XML in the fashion of MongoDB or MarkLogic, or more generally described using a kind of DDL in something like Apache Cassandra.
There are languages like UML that try to describe conceptual schemas, but these aren't widely used these days, although they were a "hot thing" for a time in the mid and late 1990s. For the most part in the relational world, the closest thing to a "language" for a "conceptual schema" is probably Entity–relationship model diagramming and description tools. Tools like Embarcadero Technologies E/R Studio can reverse-engineer a ER diagram from an existing SQL schema and generate SQL DDL from a user-created ER diagram.
The "view level" or "external level" is typically implemented outside of data managers in user-side code, possibly using Object-relational mapping tools like Hibernateto convert structures that are in "physical layer" schemas to more "external-friendly" structures that ultimately map to display or user input UI elements in application code. Other code may just have a "database layer" that completely implements this interface inside the API (and a lot of code will use both a "database layer" API as well as OR mapper tools and libraries).
(Frankly, I don't think database views really are "view layer" things, as they are still quite "physical" in their implementation and behavior. But others may disagree...)
Similar questions