Data Conversion - Information Systems Reengineering, Integration and Normalization - page 158

The pseudo code for this process can be described as follows:

Begin

Get all relation R 1 , R 2 ….R n within relational schema;
For i = 1 to n do /* first substep: load each classwith
corresponding relation tuple data */
begin
while R j tuple found do
output non-foreign keyattributevaluetoasequential file F i
with insert statement;

end;

For j = 1 to n do /* second substep: update each loaded class with
its association attribute value */
begin
while R j tuple with a non-null foreign key value found do
begin
Get the referred parent relation tuple fromR p whereR p is
a parent relation to R j;
Output the referred parent relation tuple to a sequential
file F j with update statement;
Get the referred child relation tuple from R j ;
Output the referred child relation tuple to the same file
F j with update statement;

end;

For k=1tondo/*third substep; updateeachsubclassto
inherit itssuperclass attribute value */
begin
whileasubclass relationR k tuplefound do
begin
Getthe referred superclassrelation tuple fromR s
where R s is asuperclassrelation to R k ;
Output the referred superclass relation tuple to a
sequential file F k with update statement;

end;
end;

Step 2—(Optional) Transfer sequential files to target computer.
The unloaded sequential file can be transferred to another computer if the target
object-oriented database resides on another machine. The data format may need to
be changed due to different bit size per word and/or character size per record. This
is a straightforward task for which many software utilities already exist.