One of my customers is thinking about how to organise regression test. Goal is to run automatically a script for 2 models (on older version and the current version with new changes) and to see if the calculated results are the same. This is to prevent having full integrated each time when changes have applied.
Question is: does any of our customers already have such a solution in place? I have been searching for it but it doesn't seem to be the case. So this is my last option to get an answer.
Hi Johan, could you not do this with Anaplan Connect and Versions?
Step 1: Create a separate regression model with import processes to bring in values from the source model before any changes were made. These processes would import the data into a pre change version.
Step 2: Create a second set of processes to import data from the source model. These processes would import the data into a second, post change version.
Step 3: Create a third version used to calculate the variance between the Pre and Post change versions. Dashboards could then be set up at summary and detail levels to identify where any variances between the two versions occur.
I assume we would only need Output data to be present to validate the changes and no calculations would be required so this should limit the size of the model for the customer. Additionally if Anaplan connect was used processes could be scheduled to run as required once changes had been pushed out, or triggered manually by triggering the processes at the required point.