How to use timeout in plsql

Question

Hi,
I have one procedure which calls a function. My requirement is if that function didnt return value in 1 minute then stop the exection of that function and call next statement.
Begin
(1).........
(2).......
(3) get_func_val:= function(1,2);
(4)......
(5)....
end;
My requirement is, if (3) doesnt return value in 1 minute, stop the execution of (3) and start executing (4).
Regards,
Parag

Answer 1

ParagNarkhede wrote:
My requirement is, if (3) doesnt return value in 1 minute, stop the execution of (3) and start executing (4).
PL/SQL does not support this as a language feature. It does not support a threading or fibre process model.
Parallel processing can be done via DBMS_PARALLEL_EXECUTE - but this is still not like the threading models on C/C++, C#, Java, etc.
One a parallel process is started, the main process cannot interrupt it (there are no suspend or terminate thread type interfaces to call). So the main PL/SQL process can itself timeout waiting for the parallel process (e.g. DBMS_JOB or DBMS_SCHEDULER background process). But it cannot timeout and halt the background process.
As the PL/SQL model is process based and not thread based, the parallel process does not share the main process's memory and data (or inherit a copy of it like a fork() child process would).
As for IPC interfaces. Standard concurrent access to a SQL table, serving as shared memory, can be used. Messages can also be send via the DBMS_PIPE and Advance Message Queue interfaces.
There are numerous reasons for the way PL/SQL approaches the concept of threading/parallel processing. Prime amongst these is that this is a multi-user and multi-process server environment  - able to service a multiple threaded clients. Pushing the client thread concept into this environment does not make much sense.
As for your problem. Why does step 4 need to wait for step 3, if step 4 can be executed anyway when step 3 is timed out (i.e. step 3 fails)?
If step 4 can continue without step 3 completing, then step 4 does not have a dependency on step 3. In which case the following process model can easily be used in PL/SQL:
main process()
  exec step 1
  exec step 2
  start step 3 as a background process
  exec step 4
  wait for step 3 to complete if not already completed
end
Perhaps if you explain what step 3 is, what it is tasked doing, and why it can be timed out, forum members can provide concrete solutions and suggestions.