Knowledge Management

Question:
Hi all,
I have one program which fetches data for orders and sales and generates one output file. I have order date as select option. The program once reads all the required orders and sales tables for a given date range and all the data are stored in internal tables. Now it only loops thro' all the internal tables and does some processing. IF i run the prog to fetch full year data, it takes a long long time. So i have to divide it in 3months each. S
I am scheduling the same program with 4 different job. Each job running for 3 months duration (order_date = jan to mar, apr to june etc).
Everything goes fine but sometimes , the job gets hanged in D020L table.
Now basis ppl have given me OSS note for this.
--------------------
Solution as per an OSS note is as follows
You can avoid this symptom by executing the report once online before performing these simultaneous batch runs.The selection screen then already exists when the batch runs are started.Similarly, you should avoid scheduling parallel batch jobs if changeswere implemented that result in generation of the corresponding report.
Another option is to reschedule the attached report as job step, which is then to be scheduled before the actual batch job.
Using the selection criterion s_devcl for the development class and s_prog for the program names, you can select the programs for which the ABAP load and the screen load is created before the actual batch jobs.Generating the ABAP and screen load "on demand" is no longer necessary.No blocks are necessary on the D020L in this case.
--------------------
I am not able to understand what to do exactly. i am not clear in the last option (Using the selection criterion s_devcl for the development class and s_prog for the program names ).
Can you pls help and explain me in detail??
Answer:
Pls pls help...ITs urgent
Answer:
Now you have 3 jobs,each job process for 4 months data. Did you try to schedule the jobs sequentially, try doing that. Do not run those jobs paralell.
Answer:
i need to run these jobs parallely as after all the jobs are over, i am running one more job which combine output file generated by each of these 4 jobs and this full activity should be over in 4-5 hrs.
If i schedule jobs sequentially, the total time will be around 13-14 hrs.. which is not acceptable.
Pls help
Thanks,
Veena
Answer:
i need to run these jobs parallely as after all the jobs are over, i am running one more job which combine output file generated by each of these 4 jobs and this full activity should be over in 4-5 hrs.
If i schedule jobs sequentially, the total time will be around 13-14 hrs.. which is not acceptable.
Pls help
Thanks,
Veena
Maybe it is sufficient to schedule the jobs almost parallel (in a delay of ca 5 min) since generating causes the problems.
On a general note you should consider if other possibilities could improve your performance issues. (Transferring summarized data every month to avoid rereading of the original data if the whole year is selected for example).
Christian