Check Conditions and exceptions

Question

I have the two tables:
The first is a Chequing accounts table. This table has two
number fields, the first is called balance, and the second is
called Overdraft.
The second is a Savings accounts table. This table has one
field of importance, and that is the balance alone.
I was thinking of making triggers that would fire after the
transactions are executed. If the trigger finds that the either
the Chequing balance is lesser than the negative overdraft, or
if the Savings balance is lesser than 0, then the trigger would
prevent the changes from being made.
I was considering the use of a trigger that would spread an
exception to the Transaction procedure. This would work as
follows (Pardon me if the code has problems, I'm just writing
from scratch):
create trigger chq_acct after update on chq_acct for each row
declare
declare exception insufficient_funds here
begin
if balance < (- overdraft) then
raise insuffiecient_funds
end
The things I need to know is how to pass this exception onto the
procedure that is handling the update. However, I have not seen
examples on how this is done yet.
Another possibility is more modest and probably easier to
handle: using a check constraint to verify data. In the savings
acct, I can simply add a check constraint and if it is violated
during update, catch the exception with the appropriate code
(which I don't know what it is). However, I do not know about
the check constraint for the Chequing accounts. It should be
Balance > (- Overdraft), but this is a "variable" constraint as
different customers have different levels of overdraft. Does
Oracle handle this kind of constraint? If not, how should one
handle it?
Thanks

Answer 1

hi....
Trigger is a good idea.
You can raise an application error in the trigger. Carefully
define the two constants, ...msg and ...num - especially since
the num should not be used for other exceptions. I usually put
them in the specification of an PL/SQL block for global access.
I added them to your code an example of what's needed.
create trigger chq_acct after update on chq_acct for each row
declare
kc_sqlerrmsg_insufficientfunds CONSTANT VARCHAR2
() := 'Insufficient funds.' ;
kn_sqlerrmum_insufficientfunds CONSTANT NUMBER(5,0) := -
20000 ; -- between -20000 and -20999
begin
if balance < (- overdraft) then
--raise insuffiecient_funds
     RAISE_APPLICATION_ERROR ( kn_sqlerrmum_insufficientfunds
kc_sqlerrmsg_insufficientfunds
end Then when ever an update throws the exception, the query will be
rolled back UNLESS you handle your new exception.
Put your update query into an PL/SQL block with the exception.
Be sure to tie the exception in the trigger using your defined
exception and the all important ...num.
The WHEN OTHERS THEN RAISE allows other exceptions to be handled
by other, containg PL/SQL blocks.
In the EXCEPTION part of the PL/SQL block, when the
insuficient_funds exception is thrown, the variable SQLCODE will
have your ...num in it and SQLERRM will have your ...msg in it.
This is cool because when another UPDATE not in a block throws
the exception in the trigger, whoever is trying to figure out
what happened can use these!
DECLARE
  kn_sqlerrmum_insufficientfunds CONSTANT NUMBER(5,0) := -
20000 ; -- between -20000 and -20999
     insufficient_funds     EXCEPTION ;
     PRAGMA EXCEPTION_INIT ( insufficient_funds,
kn_sqlerrmum_insufficientfunds )
BEGIN
     UPDATE  .... SET balance .... ;
EXCEPTION
     WHEN insufficient_funds THEN
      ...code to handle the potential overdraft...
     WHEN OTHERS THEN RAISE ;
END ;
Hope this helps. Good Luck.