How to create a schedule to run once?

Question

I am having some problems getting a schedule to run only once. If I leave the repeat_interval empty, the schedule does not run at all. The documentation explains all the options for setting the repeat frequency, but it does not explain how to run it only once at a set time/date. I have created a user interface to allow the users to create their own schedules and I'd rather not do something like a "yearly" interval with an end date of next week. I need the generic solution.
Any ideas?

Answer 1

Hi Ravi,
Thanks for your reply!
I have just discovered something very interesting! best explained by example:
(I might log an SR through metalink too. I'll be sure to update this thread with the outcome...)
I create 3 schedules and 3 simple jobs:
schedule 1 - start_date in 2 minutes time (eg. 11:30:00), repeat_interval => null
schedule 2 - start_date in 2 minutes, 20 seconds time (eg. 11:30:20), repeat_interval => 'FREQ=DAILY'
schedule 3 - start_date in 2 minutes, 40 seconds time (eg. 11:30:40), repeat_interval => null
jobs 1, 2 and 3 just insert a row into a table through a PL/SQL block.
My test results are as follows:
1. The first time these schedules are created only schedule 2 runs. Schedules 1 and 3 never run.
2. I drop and re-create all schedules and jobs, (adjusting the time to a few minutes in the future) and re-submit the schedules and jobs and they all run. I can continue this as many times as I like now and the repeat_interval of null will now work fine.
3. re-start the 10gR2 database.
4. re-create the schedules and jobs and only schedule 2 runs again. Schedules 1 and 3 never run. (I'm back at step 1!! I can reproduce this problem anytime by following these steps)
My script is below:
declare
start_datetime__w timestamp(3) with time zone;
repeat_interval__w varchar2(200);
begin
* REPEAT INTERVAL NULL TEST
* TIM_SCHED_2_1 - no repeat interval
* TIM_SCHED_2_2 - a repeat interval
* TIM_SCHED_2_3 - no repeat interval (identical to TIM_SCHED_2_1)
* Results on my 10gR2 database (Redhat Enterprise Linux 4.4) are:
* job 2 will always run
* jobs 1 and 3 will only run after the successful run of job2 AND jobs 1 and 3 have been re-created
* Why?
* It seems the scheduler needs a wakeup call!!!
* test table: create table tim (col1 varchar2(30), col2 date);
-- Cleanup previous jobs:
--dbms_scheduler.drop_job      (job_name      => 'TIM_JOB_2_1');
--dbms_scheduler.drop_schedule (schedule_name => 'TIM_SCHED_2_1');
--dbms_scheduler.drop_job      (job_name      => 'TIM_JOB_2_2');
--dbms_scheduler.drop_schedule (schedule_name => 'TIM_SCHED_2_2');
--dbms_scheduler.drop_job      (job_name      => 'TIM_JOB_2_3');
--dbms_scheduler.drop_schedule (schedule_name => 'TIM_SCHED_2_3');
repeat_interval__w := 'FREQ=DAILY';
-- create schedule TIM_SCHED_2_1, repeat_interval__w = null
start_datetime__w := '27-FEB-2007 11:34:00.000 AM +10:00';
dbms_scheduler.create_schedule (schedule_name => 'TIM_SCHED_2_1',
start_date => start_datetime__w,
repeat_interval => null,
end_date => null,
comments => 'schedule created at '||to_char(sysdate,'DD-MON-YYYY HH24:MI:SS'));
-- create schedule TIM_SCHED_2_2, with a repeat_interval set
start_datetime__w := '27-FEB-2007 11:34:20.000 AM +10:00';
dbms_scheduler.create_schedule (schedule_name => 'TIM_SCHED_2_2',
start_date => start_datetime__w,
repeat_interval => repeat_interval__w,
end_date => null,
comments => 'schedule created at '||to_char(sysdate,'DD-MON-YYYY HH24:MI:SS'));
-- create schedule TIM_SCHED_2_3, repeat_interval__w = null (SCHEDULE IDENTICAL TO TIM_SCHED_2_1)
start_datetime__w := '27-FEB-2007 11:34:40.000 AM +10:00';
dbms_scheduler.create_schedule (schedule_name => 'TIM_SCHED_2_3',
start_date => start_datetime__w,
repeat_interval => null,
end_date => null,
comments => 'schedule created at '||to_char(sysdate,'DD-MON-YYYY HH24:MI:SS'));
-- create jobs
dbms_scheduler.create_job (job_name => 'TIM_JOB_2_1',
schedule_name => 'TIM_SCHED_2_1',
job_type => 'PLSQL_BLOCK',
job_action => 'begin '||
'insert into tim values (''TIM_JOB_2_1'', sysdate); '||
'commit; '||
'end;',
number_of_arguments => 0,
enabled => TRUE,
comments => 'job created at '||to_char(sysdate,'DD-MON-YY HH24:MI:SS'));
dbms_scheduler.create_job (job_name => 'TIM_JOB_2_2',
schedule_name => 'TIM_SCHED_2_2',
job_type => 'PLSQL_BLOCK',
job_action => 'begin '||
'insert into tim values (''TIM_JOB_2_2'', sysdate); '||
'commit; '||
'end;',
number_of_arguments => 0,
enabled => TRUE,
comments => 'job created at '||to_char(sysdate,'DD-MON-YY HH24:MI:SS'));
dbms_scheduler.create_job (job_name => 'TIM_JOB_2_3',
schedule_name => 'TIM_SCHED_2_3',
job_type => 'PLSQL_BLOCK',
job_action => 'begin '||
'insert into tim values (''TIM_JOB_2_3'', sysdate); '||
'commit; '||
'end;',
number_of_arguments => 0,
enabled => TRUE,
comments => 'job created at '||to_char(sysdate,'DD-MON-YY HH24:MI:SS'));
end;
Cheers,
Tim.