Using Partition IDs

When you have a large volume of transactions to process in a limited run-time, use partition IDs to enable the journal posting processes to run in parallel. Partition IDs enable you to group your transaction data into mutually exclusive sets, ending contention between processes for the same row of data during posting. This eliminates possible delays due to halted processes and database rollbacks.

Only after your system has been in production for several months is it practical to implement partition IDs. It requires reevaluation at regular periods to ensure your system stays within the limits of your batch run-time window.

It is crucial that you understand your data structure before attempting to use partition IDs. Analyze the volume of your transactions by business unit and ChartField, and divide them into transaction groups with roughly equal numbers of transactions. Using your analysis, create a business unit/ChartField combination scheme to classify the transactions into mutually exclusive data sets. Then, create a partition ID for each a business unit/ChartField combination.

You might want to monitor the duration of batch processing to ensure your partition IDs remain effective over time. This can provide advance warning of changes in your system's batch processing requirements, as well as changes in your business.

To use partition IDs, first, define them using the Partition ChartField Entry Process Partition page. Then, specify your partition IDs in the request pages for the Journal Edit and Journal Post processes.