Data model sections are only referred to by query sections. Therefore, as long as the new data model can be attached to the query sections correctly, the rest of the document continues to work as expected.
The coupling between a query section and a data model section is through symbolic references, based on the names for the tables and columns exposed by the data model section. A small number of more complex dependencies exist regarding filters, however coupling basically relies on names. If two data models expose the names required by a query; for example, those names used in the Request and Filter lines, then either data model can support that query. If data model A exposes equivalent or more names than data model B, A is a valid replacement for B.
The concept of logical names is vital. Data model sections translate database table and column names into logical names. The generated names are the database names with spaces replacing underscores. The first letter of each word is capitalized by a default, however, the logical names can be changed by users. If a table is used twice in a data model, the logical name generated for the second instance appends a number. Therefore, if the Dept table is displayed twice in a data model, the second instance is called Dept2. The names referenced in a query are always the logical names of the data model.
The Update Data Models transformation leverages symbolic coupling, by using the logical names that are independent of the database names to perform tasks.