SQL & PL/SQL

Hello,

I am interested in knowing proper usage of 'savepoint' within a PL/SQL control block. Between the dashed lines below is the general circumstance (please ignore the syntax inaccuracy). My questions are:

a) Assuming the loop does NOT encounter the Exception, will repetitive iterations of the PL/SQL runtime engine seeing the 'Savepoint marker(x)' have an adverse affect?

b) I am fairly sure that the COMMIT is necessary within the Exception section to secure the effect of the 'Rollback' statement. If so, does this affect the cursor status?

Ideally, I would like to maintain the integrity of the cursor, and have the DML statements be autonomous. Can this be accomplished within (1) control block, or do I need to create a separate PL/SQL segment for the DML section to achieve this goal?

Thank you..


----------------------------------------------------------------------------------------------------------------
Begin

Open cursor for update

Loop until NOTFOUND

< Some Non-DML processing events occur>....

Savepoint marker(x)

<DML statement to Insert>
Exception
Rollback to marker(x)
Commit (main part of question)

<More processing events>

End Loop

Close Cursor

Commit

End
-----------------------------------------------------------------------------------------------------------------