Data Generation vs Extraction

As you are missing test data, you would be tempted to obtain those data by extracting subsets of your production data. That sounds good, you will be able to run your test cases on "real" data. In fact, extraction approach leads to some hidden difficulties that can be avoided when using a generation tool like GEDIS Studio.

The extraction burden

Extraction-based test data creation is a long and costly process:

1. You will need to specify your test needs, turn it into SQL requests and extract the subset of data that is of interest for your test.
2. You will need to observe that the extracted data does not fulfil your needs, think about what is missing and write some SQL requests to duplicate and slightly transform some data rows
3. You will need to anonymize every single sensitive data, which is a complex problem by itself (see here).
4. You will need to generate some additional data to test brand new features introduced in the tested release
5. You will need to redo these tedious operations for every new test campaign

The rational behind those difficulties is that extraction tools make you focus on the databases structures rather than on the data you really need.

Generating realistic and relevant test data

Actually, what you need is not real data but realistic data adapted for a specific testing objective.

Data generation is the most natural alternative to the tedious data extraction process: you know what the test data you need look like, why don't you just produce them ?

Data generation with GEDIS Studio contains only two steps:

• data models design: you describe what the data are made of (e.g. an ATM id, a last name,...) and how the different fields are correlated (e.g. the date of employment is at least 20 years after the birth date)
• test data production: you configure the data generator to obtain contrete data rows that fulfil a given testing objective

Consequently, you can design data models that are common for different applications and test campaigns. For each new test campaign, you only focus on testing objectives and configure the data generators. Doing so, you obtain unlimited amounts of realistic and relevant data while keeping your production data safe !

Follow the links to discover how GEDIS Studio can produce business specific test data for you: ATM transactions, Call Detail Records, Shopping Carts