The plan update process reads and reverses data into a planning pending file based on dimensions, categories, and values selected using the client software. Once the data is pending, the process loads data into warehouse data tables.
To improve the performance for the plan update processing data into the pending file, we recommend creating an index.
- If the plan update is without a breakdown, create the index over the low-level header table based on the dimensions and category selected for the plan update (i.e.BD999041).
- If the plan update is with a breakdown, create the index over the low-level header table based on the dimensions and category selected for the plan update (i.e.BD999041). You must create another index on the dimension and category selected for breakdown category. The breakdown process uses the best-fit DSL (Data Summary Level).
To improve the process loading data from the pending file, we recommend running the Plan update job on a dedicated memory server.
- For iSeries Server: The WRTBCK command is called and should be run in a single threaded subsystem with the maximum amount of memory. The more memory allocated to the subsystem, the faster the load process will run. Please see our documentation on creating a dedicated subsystem.
- For Windows Server: The batch file WRTBCKB.BAT runs the PLDSPPLN.EXE and should run with maximum amount of memory dynamically configured for SQL server. The more memory configured for the server, the faster the load process will run.