Our customer organization identified a problem with one of our applications. So they submitted a high priority trouble ticket. Internal testers working for our company could not duplicate the problem in their test environment. The development team lead called me asking why the customer made this a high priority problem. I told him I could think of many reasons why they would justify such a classification. However I recommended we just fix the problem and not try to fight whether it was high priority or not. Due to my recommendation, the problem got assigned to me.I always like to do a little homework on the problem assigned to me before I call the customer. So I installed the application transmittal they were running. However I could not make the problem happen. I put a call in with the customer that opened the problem. She was not at her desk. Due to the high priority nature of the problem, I tracked her down in her test lab. She walked me through the steps she was taking. That’s when I figured the problem must be related to the data in her test database. It was missing some values.
This should have been no problem. I got busy writing a PL/SQL script to correct the data in her test database. I just needed to insert a bunch of rows in one table. At first I thought I wanted a bunch of nested loops to control the values I inserted into each of the columns. However they were character data and I could not figure out how to look through some char values. Shame on me. Time was ticking away on this high priority problem.
In the end, I decided to get down to business. I wrote a small helper nested procedure to do the dirty work. I hard coded some values in it. Then I hard coded a bunch of calls to the helper procedure. It was not clean. But it got the job done. I made sure to order the parameters to the procedure so that calls to it at least look good in the source code. I did use one numeric loop to control some of the calls to the local procedure. But I needed to format and convert the numbers to text before making the call.
While unit testing I found that my script generated some exceptions. I used the tried and true method of commenting everything out until I knew what the problem was. More unit testing showed that I got some of the values wrongs. I fixed these. Then I also wrote some SQL for our own testers to run in order to recreate the problem in their database. I think I did a pretty good job with this trouble ticket. From the moment I got the call to work this problem,. It was well under 2 hours before a solution was delivered to test.
