Generator frameworks
A generator framework defines the method used to produce synthetic data values. When creating a generator, you select a framework on the first step of the Create Generator wizard. The framework you choose determines how the generator produces output and what configuration options are available in the subsequent steps. For more information on creating a generator, see Create a generator.
The following frameworks are available in Synthetic Data.
-
Generates Date and Time values.
-
Multi-Output Business Rule Generator
A version of Business Rule Generator that produces several named output columns from a single invocation. The rule body returns a dictionary keyed by each output's declared name.
-
This generator generates fixed values supplied by the user.
-
This generator takes in seed lists and generates values from them.
-
Generates Timestamp values.
-
Generates data for multiple dataclasses using datafaker.
-
This generator generates unique sequential values with optional formatting.
-
Generates values by choosing one of several other generators using a weighted distribution method.
-
This generator generates UUID values.
-
This generator generates data based on regex supplied.
-
Looks up a value from another table, such as finding a product's price using the product ID on the current row.
-
Generates data based on references. Configure this framework at the dataset-field level to draw generated values from related data and maintain referential consistency.
-
A customer- or LLM-authored per-field rule that produces a column's values, either once per row or once per batch for vectorized output.
Null Value Generator is a built-in, system-provided generator. DCT provides one preloaded instance per framework. You can assign it to any field, and discovery can assign it automatically. It generates a database NULL value, not the string null or NULL. You cannot create, configure, or customize this generator, and it is excluded from the generator framework list and custom generator workflows.