Test data is the input and supporting information used to trigger software behavior and observe results. Realistic, useful, anonymized data can determine whether testing succeeds, but creating and managing it consumes substantial time and resources. section 1.1.1 section 6.0
LLMs can rapidly generate, transform, and expand test data, provided that prompts state the required structure, values, relationships, format, and validation rules explicitly. Their output remains probabilistic and requires human checking. section 6.1 section 6.4
Useful practices include:
- Define fields with delimiters, data types, allowed values, ranges, and randomization rules. section 6.1.1
- Specify the output format and ask the generator to check record counts, types, syntax, and relationships before returning data. section 6.1.1 section 8.2.2
- Inspect generated values against the system’s actual validation rules; plausible values may still be invalid or semantically inconsistent. section 6.1.1 section 7.2.1
- Protect personal and organizational privacy when supplying examples or existing data to an LLM. section 6.4
Generating structured room data
A prompt can request five JSON objects and define room_name, type, beds, accessibility, image, description, features, and roomPrice with explicit types and permitted values. The resulting records may look valid, yet example.com image URLs may not identify real images and beds may not match the room type. Tightening the prompt and validating the result against application rules makes the data more useful. section 6.1.1
Seeding large datasets
For exploratory testing requiring thousands of bookings, an LLM can expand an existing INSERT statement into SQL seed data for rooms and bookings. Prompt rules may need adjustment for array fields, while a script can generate 1,000 bookings with randomized rooms, dates, names, and deposit values. section 8.2.2
% room_name | string | random % type | string | 'single' or 'double' % beds | integer | 1 to 6 % accessible | boolean | true or false % image | string | random url % description | string | random max 20 characters % features | array[string] | 'Wifi', 'TV' or 'Safe' % roomPrice | integer | 100 to 200
Shows a compact delimited specification for generating room records, including field types, ranges, allowed values, and randomization instructions. section 6.1.1
