Why Test a Database?
Just as developers are human and prone to errors, a database containing mission critical information and lines of code on which applications are built is also vulnerable to errors and requires regular maintenance and updates from time to time. Organizations use several applications day in and day out that work on a database which get altered depending on the requirements, which affect database testing. The best way to ensure the functioning of these applications is to test them on a regular basis for new features and changes made in the database.
This is what regression testing is all about.
What is Regression Testing of a Database?
Regression Testing of Database refers to testing the database for any new changes and updates made in the database to keep its integrity intact even after modifications in it.
Steps for Database Regression Testing
Step 1: Decide What to Test in Database Regression Testing?
When it comes to performing regression testing of a database, it is never pre-defined as to which part or kind of database gets updated that requires retesting. The updates keep on occurring in various areas of database and testers need to target those updates to get the most out of their efforts. Generally, database testing comprises of following:
- Incoming data values,
- Outgoing data values (generated for raw queries)
- Database elements (tables, procedures)
- Metadata for web applications
- Data load that’s somewhat we call "load testing"
While performing regression testing of a database, if you are adopting an approach of either Black-box testing or White-box testing, your contents for testing would be something like this:
- Black-box Testing at the interface
- O/R mappings (including the metadata)
- Incoming data values
- Outgoing data values (from queries, stored functions, views ...)
- Stored procedures & functions
- White/Clear-Box Testing internally within the database
- Database Schema (tables, procedures, etc.)
- View constraints
- View definitions
- Referential integrity (RI) rules
- Default values for a column
- Data invariants involving several columns
Step 2: Decide When to Test Database for Regression Testing?
When you change something in the existing database or application that runs on the database to refactor it or add a new functionality, you need to ensure that you have not broken anything. This requires you to perform regression testing to either fix the errors or roll back the changes made in the database. Hence, regression testing is generally done after introducing changes in the database to check the behavior of the modified code. It requires hundreds of new tests to retest in order to check and rectify the newly written code as well as integration with other systems or applications.
If you are working in an agile environment, you need to take the Test-First Approach or Test-First Development (TFD). In this approach of regression testing, you need to prepare test code parallel to development in order to test the business logic implemented in the database as well as forms, data validation rules, referential integrity, etc. in the database which works like a cycle as mentioned below:
- Add a test
- Run your tests
- Update the code or make changes
- Run your tests again
Step 3: Decide How to Perform Regression Testing of Database?
This is the most important question that comes to the testers’ mind while performing regression testing of the database. Regression testing can be performed both in a manual and automated manner.
A common practice in an agile team is for testers to have their own "sandboxes" with which to work. A sandbox is basically a technical environment that isolates untested codes from the production repository. You can opt for either of them depending on the size of the tests and database.
In my opinion, automated regression testing is more beneficial since it allows testers to test and manage a large number of databases without manual intervention. In other words, it allows you to perform other tasks while the test tool continues to do its job in the background and sends the report of bugs if any.
Automated Regression Testing of a database seems easier and more convenient, while adding to the testers’ productivity by saving a huge amount of time along with imparting the following benefits:
- Entire database regression testing process gets automated
- Reduces testing efforts by 50% to 75%
- Testing is done parallel to development
- Extensive test coverage can be achieved
There are several test automation tools that help perform regression testing. These automation tools help in getting the testing team up to speed with the database regression testing requirements in no time, overcoming several challenges they face in manual testing.