You load a set of employees into your test instance, start validating the data and then spot an issue. The conversion file has used the wrong values, duplicate test records have been created, or the next testing cycle needs a fresh set of employees.
At this point, correcting individual transactions may take longer than preparing the data again. Can you remove the affected employees and start fresh?
Yes. Oracle provides the scheduled process Purge Person Data in Test Environments for this purpose. It is enabled by default in Oracle-supported Test/Stage environments and cannot run in Production. Oracle process overview.
Let’s look at the selection options, the preview check and the clean-up activities that make this useful during implementation.
Why Would You Need to Purge Test Employees?
Think about a conversion cycle where a small group of employees has been loaded incorrectly. You want to fix the source file and repeat the test without mixing the earlier records with the corrected load.
The same approach can help when a demonstration has finished or a defined test population needs to be removed before the next round of testing. Keep the population specific to the records you intend to remove.
What Does the Process Remove?
The scope extends beyond the employee’s employment record. Oracle lists person details and dependent HCM data, including assignments, work relationships, absences, benefits, calculation cards, element entries, compensation, documents, goals, performance records, talent profiles, and roles and user details.
Oracle also publishes the underlying tables covered by the process. Examples include:
| Area | Examples from Oracle’s table list |
|---|---|
| Person and employment | PER_ALL_PEOPLE_F, PER_ALL_ASSIGNMENTS_M, PER_PERIODS_OF_SERVICE |
| Documents and Journeys | HR_DOCUMENTS_OF_RECORD, PER_ALLOCATED_CHECKLISTS, PER_ALLOCATED_TASKS |
| Absence | ANC_PER_ABS_ENTRIES |
| Payroll | PAY_ELEMENT_ENTRIES_F, PAY_PAY_RELATIONSHIPS_F |
| User information | PER_USERS, PER_USER_ROLES |
See Oracle’s complete list of tables for the detailed coverage. These tables describe the scheduled process’s scope; they are not instructions for deleting data directly.
Before You Start
Confirm the Test/Stage instance, the exact employee population and the person numbers you plan to use. Retain the source files and any test evidence you need for comparison after the corrected load.
The user running the process requires the function security privilege HRC_PURGE_PERSON_PROCESS_PRIV. Oracle execution instructions.
If the process is unavailable in your Test/Stage instance, check your access. Oracle’s availability guidance advises raising an Oracle Support service request when the process cannot be found in a Test/Stage environment.
Preview the Population Before You Purge
Use the preview as a deliberate checkpoint. Compare the returned population with your intended employee list, including the count and identifiers. A broadly matching selection can include records you did not intend to remove.
Step-by-Step: Purge the Selected Employees
1. 🔎 Open the Scheduled Process
Open Navigator, select Tools, then Scheduled Processes. Choose Schedule New Process and select Purge Person Data in Test Environments.
2. 🎯 Identify the Population
Use one documented selection method:
| Parameter | What to provide |
|---|---|
| Person ID | Internal IDs, separated by commas for multiple people. |
| Person Number | Person numbers, separated by commas. Wildcard patterns are supported. |
| Person Query | A SQL query that returns the person IDs to select. |
For this walkthrough, use a small, explicit list of test person numbers, such as TEST1001,TEST1002. Keep the other selection fields blank.
3. 👀 Submit the Preview
Set Save = N and submit. This lists the selected people without purging their data.
Review the request’s available output and log against your intended employee list. If the selection is incorrect, refine it and preview again.
4. ⚠️ Check the Payroll Restriction
Oracle states that a worker with processed payroll actions is not purged. The relevant payroll processes must first be rolled back. Work with the payroll team to establish the appropriate rollback sequence for the affected test workers. Oracle restriction.
Oracle’s Fusion Centre of Excellence guidance also highlights payroll history and balance loads when discussing this restriction.
5. 🗑️ Run the Actual Purge
Submit the same verified selection with Save = Y. This performs the purge. Oracle execution instructions.
6. ⏳ Allow Clean-up to Finish
Allow the process to complete. Oracle advises against cancelling it because its final stages perform clean-up; interrupted clean-up resumes when the process next runs.
7. 🔐 Complete the LDAP Processing
The purge generates user-account deletion requests. Send Pending LDAP Requests processes those requests after the purge completes. If an existing schedule handles them, verify that run; otherwise submit the process.
Use User Type = All and Batch Size = A. These are the settings in Oracle’s purge instructions, with A calculating the batch size automatically. Oracle LDAP process guidance.
8. ✅ Validate the Outcome
Compare the outcome with the original employee list. Check the purge log for skipped records or errors, verify person removal using appropriately authorised access, and check the associated user accounts after LDAP processing.
A person disappearing from one search page is insufficient evidence on its own. Record the process results and checks together, particularly before a reload.
The Complete Process at a Glance
Use this image as the reference when working through the activity. Treat the person-data result and user-account result as two checks to complete before closing the test clean-up.
What Is Excluded?
Oracle excludes predefined person data, contacts enrolled in a Benefits programme and Trading Community Architecture (TCA) data. Dependent TCA records are inactivated rather than purged. The processed-payroll restriction also applies. Oracle exclusions.
Keep these boundaries in mind when validating the result. Avoid describing the process as removing every trace of a person from every connected system.
Is This the Same as Cancel Work Relationship?
No. Cancel Work Relationship removes the work relationship and associated employment data, while retaining the person record. Oracle notes that a person whose only work relationship has been cancelled can be available for rehire even when excluded from general person searches. Oracle Cancel Work Relationships guidance.
This distinction matters during testing. Choose the action based on what you need to correct, rather than assuming that cancelling employment deletes the person.
Practical Use Case: An Incorrect HDL Conversion
Imagine that two test employees were created from a file containing incorrect values. The team decides to repeat the conversion for those employees.
First, compare the preview with those two intended records. Complete the purge and LDAP checks, correct the source data and HDL file, then perform the fresh load. Validate the new records against the expected conversion results.
Before reloading, review the identifiers and references in the corrected file. Treat it as a fresh load of the intended test population and retain the results so the next comparison is straightforward.
Final Thoughts
The next time a conversion test goes wrong, you have a supported option for removing the affected employees from your Test/Stage instance.
The useful habit is to make the selection small and reviewable, preview it before the actual purge, and finish the person and user-account checks before moving to the corrected load.
Leave a Reply