How to Compare Payroll Results Before and After a Major Schema Change
- by Owen McGivney, Senior Consultant, iProCon Ltd.
- June 1, 2011
Learn how to carry out a regression test to compare payroll results before and after a major configuration change has been applied to the payroll. Detailed steps show how to create a Microsoft Excel workbook with a simple Visual Basic for Applications (VBA) program that compares two sets of payroll results and creates a new spreadsheet that highlights and identifies all the differences. The result is a simple but effective table-comparison utility with a multitude of uses.
Visual Basic for Applications (VBA) is a programming language that enables you to programmatically carry out many of the tasks you do manually every day in Microsoft Word, Excel, and other applications. It is created automatically behind the scenes whenever you record a macro. The language and development environment are included in most versions of Word and Excel. By becoming familiar with the VBA commands and development environment you can become a super user, and learn how to automate many repetitive and tedious spreadsheet-based tasks.
Let’s concentrate on the scenario for carrying out a full-scale regression test based on payroll results. This would be the case when there has been a major change to a stable payroll system such as an upgrade or the application of Support Packages, or a significant set of changes to a payroll schema because of new laws, contractual changes, or the addition of a new corporate entity onto the system. Normally a number of unit tests are carried out to test individual items of the new functionality. However, in order to increase the level of comfort you might also want to compare the payroll results at a wage type level for all the employees on the system before and after some major configuration changes have been applied. By analyzing the results, you can ensure that no unwanted changes are inadvertently introduced.
The methodology for the regression test is fairly simple. You make a copy of the production system to create a regression test client and download payroll results for a given payroll period. You apply a set of configuration changes to the regression system, rerun the payroll, and get a new set of results for the same period. Then you compare the before-and-after results. Once you analyze the differences in the two sets of results, you can be confident that these changes are safe to go to the production client.
Would you like to see this full item?