Database Management System Emulation - Information Systems Reengineering, Integration and Normalization - page 228

Phase I. Schema translation—Map nonrelational schema to the extended entity
relationship (EER)model and then map from EER model to relational schema.
Step 1.1. Map nonrelational schema to the EER model.
1. Derive implied relationship.
2. Derive multiple (alternative) relationship.
3. Derive unary relationship.
4. Derive binary relationship.
5. Derive n-ary relationship.
6. Derive aggregation, generalization, and categorization.
7. Derive entity keys and other constraints.
Step 1.2. Map from the EER model to a relational schema.
1. Map entities into relations.
2. Map n-ary relationship into relationship relation.
3. Map aggregation, generalization, and categorization into relations.
Phase II. Relational interface—Restructure nonrelational schema and database by
adding record identifiers as secondary indices into each record type, and precom-
pile embedded SQL programs into nonrelational database programs to access the
restructured nonrelational database by using following steps.
Step 2.1. Schema restructure—Add secondary indices by record identifiers into
the nonrelational schema (from schema translation).
Step 2.2. Database restructure—Add secondary indices by record identifiers into
the nonrelational database (from data conversion).
Step 2.3. Translate embedded SQL programs into nonrelational database pro-
grams (from database program translation, i.e., logically, we are processing an em-
bedded SQL program).
Step 2.4. Process the translated nonrelational database programs to access the re-
structured nonrelational database through the nonrelational DBMS (i.e., physically,
we are processing a nonrelational database program).
Phase III. Data conversion—Convert data from a source database to a target
database by restructuring the source database sequence to the structure of the target
database sequence.
Step 3.1. Unload—The unload program applies database navigation in a pro-
gram that reads all nonrelational database records logically and unloads them with
the derived record identifiers. The record identifiers can be mapped to the relational
schema as primary keys and/or foreign keys.
Step 3.2 (optional). Transfer—Transfer unloaded sequential files from the source
computer to the target computer if they are different.
Step 3.3. Upload—The upload program/utilities can then be used to upload the
unloaded sequential files into the target relational database.
Figure 6.1 shows a diagram representation of the methodology. For more detail,
refer to Chaps. 3, 4, and 5.