Codegato Unit Testing for Crystal reports and SSRS
Unit testing is a software development practice where individual components (units) of code are tested in isolation to verify that they work as expected. The core idea is that when you test each small piece of your application independently before combining them, you can catch bugs and ensure each component functions correctly on its own. We have implemented unit testing for reports. Similarly to its function in code development you can define input parameters and an expected result (golden file) related to the parameters. The combination of input parameters and an expected result is known as a "Theory". You can periodically run the defined theories and compare the previously generated golden files with the newly exported files. Our unit testing allows you to see if the golden file and the new files are the same, and if there are differences you can analyze them fix whatever issues may arise. RT Manager is able to compare the outputs as images and as values. Comparison by image is useful when you want to detect changes in layout and content. For example, this method may detect if a field is moved to a new position. It can be useful to verify forms which are prepared for machine processing, OCR, etc. Comparison by value is useful when you want to compare Excel files and you want to ensure the values in corresponding cells are the same, regardless of the size of the columns and rows.
Sample Uses:
- Checking if the layout of the report has changed: Layout changes might be caused by new development or data changes but also by factors not directly related to the report. For example, Crystal reports uses the assigned printer page format. Changing the Default System Printer may cause changes in the Crystal reports layout.
- Verifying that reports are working in the same way on all machines: This is related to the previous point. Two computers may have different Default printers or other settings that may result in report output differences.
- Checking if the report design satisfies the rules of deterministic processing: A deterministic algorithm is an algorithm that, given a particular input, will always produce the same output. In finance, accounting, etc. it is important to have reports that show the same data for the provided period regardless of new data being added after the period.
- Detecting unexpected data changes: For this you don't even need to create a Crystal or SSRS report. You can run an Ad-hoc SQL report, export it to a Golden Excel file and later run the same report, comparing the newly generated Excel file with the Golden one. This can be used to control rates, enums, factors etc.
- Server migrations: You may need to verify that all reports are available on the new server and they are able to connect and retrieve data from the database. In addition to data and layout changes, unit testing will also catch missing report files or missing database servers.
- Analyzing expected delta: Sometimes differences between 2 runs are expected and you just want to see them. Using the Unit Test interface, you can compare 2 PDF or Excel files and easily find the differences.
- Letter proofing: When using letter templates to send to your customers, you may run a sample letter before running the batch and make sure that the letter generates the same result as the one approved during the tests. This could be useful for banks, healthcare organizations, lawyers, etc.
