Test Data Management. Part 4: generation and storage

There are strategies to get required data sets:

  • Use existing data provided by the customer without or with modification like data masking / scrambling.
  • Create test data sets manually based on random data and existing value lists: codes, names, geographic data, etc.
  • Create and run custom scripts or solutions.
  • Use the test data generation software like DTM Data Generator.

Each of mentioned way has positive and negative aspects. For example, with the second strategy, you can get test data with high quality but can't create really huge data sets.
The third way is most flexible but the solution cost can be unacceptable for most cases. As we already mentioned in the second part of the article we do not recommend to use customer's data without scrambling due to the risk of critical data compromising.

Some companies use the combination of mentioned ways to create different data sets. It seems smart strategy for complex cases and allows to create more customized data if necessary.

Our recommendations are:

  • Use the first way with data scrambling if you already have enough data sets with required properties.
  • Use the second one in case your data has a too complex format, relations or dependencies to be automated in data generation. However, test data cost may be maximal in this case due to a lot of manual operations
  • Use the third method if your data not so complex and custom scripts are easy to build. The high-quality professionals required for this way.
  • Use the last way in other cases or if you have no required resources for the first and second strategies.

The second important aspect is a storage policy for test data management components. The policy should include:

  DTM Data Generator: test data generation project
  • List of items to be stored.
  • Locations for each component of the mentioned list.
  • Version control policy for some or all items.
  • Access policy for each item.

What kind of component do we mean? There are:

  • Goals and plans.
  • Test cases and packages.
  • Libraries or predefined lists like customers, products, departments, etc.
  • Test data automation scripts or project files.
  • Test data arrays, databases, files.

Our recommendation is to share all mentioned items. The corporate server, portal like SharePoint or even cloud are good candidates. However, access policy is critical in this case to keep integrity.