Fixed multiple small issues with the new teams feature.
Comments: Fixed an issue where @mentions were not highlighted in saved comments, and where text that looked like an HTML tag was silently removed from a comment instead of being shown exactly as it was typed.
Oracle - Structural no longer copies Oracle collection columns to the destination database. Collection columns include nested tables and VARRAYs. These columns were never written correctly, and Oracle's data loading tools cannot represent them. Instead, in the destination, Structural sets these columns to NULL. The job includes a warning that names the affected table and columns. Structural returns the same warning for other Oracle column types that Structural cannot write, such as BFILE and the URI types. Previously, Structural emptied these columns without notice. A job that copies a table that contains any of these column types now finishes with the status Completed with warnings.
Fixed an issue that could cause a subsetting job to write rows whose referenced parent rows were missing, so that applying foreign key constraints failed at the end of the job. This happened when a table was reachable from a subset target through paths of different lengths, and it affected all data connectors that support subsetting.
Restored the new teams feature that was released in v1869 but reverted in v1870 because of a data migration issue.
To summarize, all organization users are added to the Default team.
On Structural Settings, the Teams tab displays the Default team. From there, organization administrators can assign team administrators.
The Teams page, displayed from the user menu at the top right, allows team administrators to view the team membership and manage team permission sets.
In addition to team management, team permissions also include the following permissions:
- Create workspaces
- Manage generator presets
- Manage sensitivity rules
- Configure secrets managers
These were previously global permissions. Workspace permission sets are also now managed within the team.
Workspaces that no longer have an owner, usually because the user was removed, are now displayed on the Workspaces Without an Owner page on Team Settings. Team administrators can then assign a new owner.
File groups: Fixed an issue where selecting files in quick succession could leave a file with no data preview and suppress the warning that the selected files do not all have the same columns.
When you delete a user who owns workspaces, you no longer have to immediately transfer those workspaces to a new owner. The workspaces are added to the Workspaces Without an Owner tab on Team Settings. From there, an admin can transfer the workspaces to a new owner or delete the workspaces. Other users that were granted access to the workspaces lose their access until the workspace is transferred.
New team permissions, permission sets, and Team Settings page
Global permissions sets are now separated into global permission sets and new team permission sets. The permissions to create workspaces, manage presets, copy workspaces, enable diagnostic logging, manage sensitivity rules, and manage secrets managers are now team permissions. The existing permissions migrate from global to team automatically.
Every organization now has a default team. Organization administrators automatically become the administrators of the organization’s default team.
The new Team Settings page, which is available to team administrators, allows them to manage team and workspace permission sets.
This is in preparation for a future enhancement to allow Enterprise organizations to create multiple teams and to manage resources separately within those teams.
You can now create custom permission sets on Structural Cloud.
Made the following changes to the Finnish Personal Identity Code generator:
The generator now preserves uniqueness, and can be used on unique and primary-key columns.
The century indicator in each generated code now mirrors the source value instead of being spread evenly across every indicator letter. This ensures that common indicators (+, -, A) remain common, and that rare reform letters only appear when the source uses them.
During data generation, Structural warns when the sampled values fall outside of the chosen birth-date range or use indicators that the range cannot represent. Structural drops those rows instead of writing them to the destination.
Aurora (PostgreSQL data connector) - Fixed an issue where PostgreSQL data generation to an Amazon Aurora destination failed on every run after the first when the source database had Aurora's pg_columnmask dynamic data masking extension installed. Structural no longer copies provider-managed extensions into the destination.
Oracle workspaces now always run on Data Pipeline V2. The option on the Confirm Generation panel to choose between classic flow and Data Pipeline V2 is removed. Existing Oracle workspaces will automatically use the Data Pipeline V2 flow.
Removed the Oracle database-link generation path and the TONIC_ORACLE_DBLINK_ENABLED environment setting. The setting no longer has any effect. You can remove it from your configuration. Database links that exist as objects in your Oracle schema are not affected. Structural still discovers them, and still warns when the destination is missing a link that is referenced by a source synonym.