Best Practices for Purging Transactions in Transaction Matching

Follow recommended best practices to obtain optimal results from purging transactions.

The right approach to archive and purge transactions is based on the following factors:

• Volume of transactions to be purged
• Shape of your data

Recommendations When Purging Transactions

• Archive matched transactions that are no longer required and then purge them.

  While it is recommended that you archive transactions before you purge them, you can also choose to purge them without archiving.

• Purge transactions outside of business hours

  When transactions are being purged for a match type, the match type is locked and operations such as importing transactions, matching transactions manually, and running Auto Match are not allowed.

• Clone your production environment to your test environment, and then purge transactions on your test environment first. The learnings from this exercise will help you purge transactions efficiently on your production environment.
• Consider rescheduling the Daily Maintenance (AMW) for that day, if the volume of transactions is large during the initial archive and purge operation. See Managing Daily Maintenance in Getting Started Guide for Administrators.

  For example, if the AMW is scheduled for 1 a.m., and the archive or purge job is still running at 11 p.m., you can reschedule the AMW to 9 p.m. of the next day. This defers the AMW to the next day ensuring that the archive or purge job is completed. However, before 9 p.m. the next day, remember to change the AMW time to its original schedule after the archive or purge job completes.

About Purging Transactions for the First Time

Depending on the size of your application and volume of transactions, the initial archival may take some time. However, once regular archive and purge routines are established, subsequent archive or purge jobs take lesser time. With larger applications, the recommendation is to perform periodic archive and purge routines. Doing this regularly minimizes the volume of data processes each time.

For optimal performance with large applications, a phased archive and purge strategy that is based on the age of transactions is recommended. This is especially applicable when working with historical data (for example, older than 18 months).

For managing historical data, a suggested strategy is as follows:

• Phase 1: Archive and purge data that is older than 18 months
• Phase 2: Archive and purge data that is older than 12 months
• Phase 3: Archive and purge data that is older than 9 months
• Phase 4: Archive and purge data that is older than 6 months

Once regular archival and purge routines are established, subsequent operations will generally take less time.

Prerequisites for Purging Transactions on the Test Instance

1. Backup your test instance. See Backing Up and Restoring an Environment Using the Maintenance Snapshot in Getting Started Guide for Administrators.
2. Clone your production instance into your test instance. See Cloning Cloud EPM Environments in Administering Migration.

Steps to Purge Transactions on the Test Instance

1. Perform a backup of your application.
2. Assess the count of matched transactions that you would purge. You can do this using the Transactions tab in Transactions Matching or through BI Publisher reports. This count can be verified after purging transactions. See Searching Transactions in Transaction Matching.
3. Select the transactions that must be archived or purged.

   Based on the volume of transactions to be archived or purged, you can specify the age of matched transactions. A single or multiple archive and purge operations can be performed. See Recommendations When Purging Transactions.

   Note: The default batch size is 200000. Do not update this setting unless recommended by Oracle Support.

4. Ensure that the Daily Maintenance process (AMW) time does not overlap with the archive or purge time. See Recommendations When Purging Transactions.
5. Purge the transactions.
6. Make note of the time taken for each step in the process.

Based on these learnings, purge transactions in your production instance.

If there are any concerns during the archive or purge operation in your test or productions instance, please create a Service Request with Oracle Support.

Purging Transactions at Regular Intervals

After you purge transactions for the first time, Oracle recommends that you purge transactions at regular intervals. You can automate this process through EPM Automate.

Steps to Purge Transactions at Regular Intervals

1. Apply the learnings from the purge process performed on the production instance.
2. Test your automated purge schedule on your test instance with the recommended settings.

   Although the recommended strategy is to archive matched transactions and then purge them, you can choose one of the following strategies:

   • Perform a simple purge

     See Performing a Simple Purge in Transaction Matching. To use EPM Automate, see purgeTmTransactions in Working with EPM Automate.

   • Archive matched transactions and then purge them.

     See Archiving Matched Transactions in Transaction Matching followed by Purging Archived Transactions in Transaction Matching. To use EPM Automate, see archiveTmTransactions and purgeArchivedTmTransactions in Working with EPM Automate.

3. Deploy the automated purge schedule on your production instance.