Release Date: September 16, 2025
Announcing targeted sampling for multi-table datasets. Use conditional sampling to provide a set of fixed values to create a targeted synthetic dataset for your project. For example, create a minimum viable test set for software testing, or rebalance a dataset. With the Targeted Sampling bundle , you can now conditional sample for multi-table datasets too, using HSA and Independent Synthesizers.
Enhancing the SelfReferentialHierarchy constraint. This constraint allows you to model a table that contains a tree-like hierarchy — for example, an org chart of employees and their managers. We’ve now enhanced the API and added support for additional types of hierarchy encodings, such as skip-levels and root nodes.