|
|
|
|
|
Control-M/Analyzer is a product designed to enhance the data balancing and quality control function within an organization. As such, it should not be exempt from its own quality control procedures.
Within a rule, consider using the quality control techniques described in this section. These techniques can be used during testing of the rule and then removed after testing. For future maintenance, however, you might also consider leaving these quality control techniques in place after testing, but modifying them temporarily so that none of their output is written to the Invocation Report.
Use totals to check that your data selections (WHEN parameters) and other DO actions are correct
You can total numeric fields within the report or file being analyzed. These total amounts can be used to check that an input file or report contains all of the data that you are expecting. This is also a way of verifying that you are, in fact, processing the correct file from the correct source and date.
Once you have established some initial totals, even on fields that you may not be verifying, you can use these as a base or control figure to compare against other balancing actions in the rule. For example, if you are extracting certain sales figures that you believe represent about 10% of the total sales in a report, you can calculate a total on the report lines extracted to see if they do represent about 10% of the total dollar amount of sales.
Have another person check any output which you are producing
When you are focusing on developing rules, you may occasionally miss some obvious errors. An important rule of quality control is to always have one person check the work of another person.
When checking the validity of a rule, use test reports or files which contain many test cases
The greater the variety of error conditions in your test report or file, the greater the chances that your rule will work as expected on the live report. On the other hand, if your test report or file is too large, you will not have the time to check everything it is doing.
After testing with your own "made-up" report or file, use a live report or file
No matter how many conditions you may think of for testing your rule, there are usually some surprises waiting for you in the real data. Before implementing your rule in production, it is a good idea to test it on some actual reports or files, and review the results with users. You may use the Control-M/Analyzer "simulation" mode for this purpose.
Keep the test reports and files, and re-use them when any changes are made to a rule. Compare output after the rule change to previous output
To ensure that you have not overlooked anything when you change a rule, you should have a "base of comparison." The results of the rule before the change should be kept on file, with sample reports and files, and sample outputs produced, including the Invocation Report.
When you make a change, all outputs should be compared.
Run the TRACE facility to debug your rule
The TRACE facility produces a complete printed output of every command executing during rule processing.
Parent Topic |