Agent support for document-based data - You can now use the Structural Agent to perform tasks related to document-based data. Document-based data includes JSON columns in relational data and document-based databases (currently MongoDB).
The supported tasks include:
- Enable and disable Document View.
- Read the assigned generators for document column and collection fields.
- Assign generators to document column and collection fields.
Note that you cannot use the Agent to create or edit generic path generators.
Structural now provides an MCP server to allow you to work with your workspaces directly from an MCP client.
Consistency for many generators, including the Name generator and the Address generator, now ignore whitespace when deciding consistency. This helps when these generators are applied to fixed width columns, such as char(40). This update does not affect other generators, such as the Character Scramble generator and Timestamp Shift generator.
Name generators that are consistent with another column now derive each part of the name (first, middle, and last) from the corresponding part of the source value. Because of this change, the output from the Name generator might differ than the output generated by earlier versions, including for workspaces that have a fixed statistics seed. Within the same version, the output is still deterministic and consistent.
v1806 was not released because of technical issues.
Snowflake - For Snowflake on AWS, you can now use an assumed role for the credentials to connect to the temporary storage location on Amazon S3.
Fixed an issue where users on the same machine and browser shared a Structural Agent chat.