After "automating" the import of some data, I need to test this. Yet I have no idea how to actually test this.

Info about the "automation":

We are having an application which has several "jobs". Each job either loads data from an external data source or calculates some values. Each day we had to select which jobs to run and then execute the jobs. As we are lazy, my job was it to implement a service, which simply executes all the jobs on a daily basis.

To be more precise, I am using the quartz-scheduler to schedule these jobs and implemented some functionality to ensure the execution order of these jobs as well as they don't run repeatedly after finishing successful.

As I tested it manually to ensure that the results are correct (compared to production system), I now need to test it officially. And this is where I have no idea on how to proceed. I need to create test documentation as well as the results etc.

So far the only meaningful thing I could come up with, is to actually test the automation and compare the results to the production system to prove that my automation yields the same results as the current system. Also I thought about appending the log file to prove that the jobs only run until they are all successful, but this does not seam feasible at all, as I basically just dump data into the test document.

So my question is, how can I feasibly test the automation of an manual process?

Please note that I am not asking about automated testing

1 Answer 1


The simplest method is to check that the results are identical. Since you are testing importing data, I'd suggest running the process manually, then exporting the data to either a text or CSV file. Next, run the process via the automation, and again, export the data to either a text file or CSV file. You may need to clear the first import before you run the second one.

After that, you can use any comparison application such as WinMerge, Beyond Compare, KDiff and check the output. If there are timestamps involved, you may need to filter the timestamps or manually check them.

I use this method all the time for testing data imports.

  • Is the simple comparison of data an accepted method to test the auomation officially? And would one need to provide the production/test data as well as the result? And how would I test that the jobs are executed automatically? Apr 12, 2019 at 12:08
  • 1
    @XtremeBaumer - The simple comparison of data is an accepted method to test that the automation generates the same data as the manual process. Whether providing the prod//test data is needed depends on the industry and regulatory requirements. Testing that jobs are executed automatically is a case of checking for the last execution time, or checking that the database has been updated with the correct timestamps, whichever works better with your process.
    – Kate Paulk
    Apr 12, 2019 at 15:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.