Database Program Translation - Information Systems Reengineering, Integration and Normalization - page 184

LET C1 BE
SELECT E.TID, E.ENO, E.ENAME, E.BIRTHYR FROM JOB, EMP
WHERE J.TITLE = JOB.TITLE AND E.ASSIGN = J.JID.
LET C2 BE
SELECT D.DNO, D.NAME, D.LOCATION FROM EMP, DEPT
WHERE E.TID = EMP.TID AND E.WORKS-IN = D.DNO.
MOVE 'ACCOUNTANT' TO TITLE IN JOB.
L.
OPEN C1.
SELECT C1.
M. IF end of set GO TO EXIT.
FETCH C1.
IF EMP.BIRTHYR ≤ 1950 GO TO N.
------ other code ------
OPEN C2.
SELECT C2.
FETCH C2.
:
CLOSE C2.
N. SELECT C1.
GO TO L.
EXIT. CLOSE C1.

5.5 Coexistence

This approach continues to support the nonrelational database while developing an
information-capacity-equivalent relational database for the same application. De-
velopers maintain an incremental mapping from the nonrelational database to the
relational database. For example, Mark et al. ( 1992 ) presented an incrementally
maintained mapping from a network to a relational database. At the beginning, the
applications on the relational database are restricted to retrievals. Gradually, appli-
cations on the network database are rewritten and moved to the relational database,
while the incremental mapping continues to maintain the relational database for the
applications still running on the network database. The basic idea of the incremental
maintained mapping is illustrated in Fig. 5.4 .
The initial network to relational database mapping algorithm takes as input the
network schema defined in terms of the network DDL. The algorithm generates an
equivalent relational schema definition in terms of the relational DDL, a program
for unloading the network database to a temporary file and a program for uploading
the temporary file to the relational database. After the relational database is defined
using the generated relational DDL statements, the network database is mapped to
the temporary file by using the generated unloading program. Finally, the uploading