Ive got an opportunity to play around with global branching recently. It really is a game changer compared to before.
But there are few drawbacks that I came across(or am I missing a setting/maybe the feature is not available in my stack just yet) -
Post transform is build in global branch, the dataset can only appears as a rawdataset in files as it is not directly possible to move around folders/files using global branch selected (I assume its “Main” by default as there is no option to make a selection)
The global branch in Ontology Manager did not make the newly created dataset available for selection when trying to create an object type (as it was not discoverable to pick using the “use existing datasource → Select datasource” which is the usual flow to create an object type)
Another drawback is when working in “Workshop” using the global branch; it does not automatically save changes
+ It will also neither detect an Action created in “main” or let us use it in the created “global branch”, so the only way out is duplicating the same
On the flip side, the best part is seeing how modified, updated, and dropped properties are handled in Ontology manager like indexing, merging and handling conflicts are perfect
Please let me know if Im missing a setting or if this is already available.
Hi, providing general guidance here based on what I could interpret from your post alone
1. Dataset placement and the “raw dataset” label
Global Branching does not give you a separate Compass folder tree. Moving a file is a shared resource-location change, not an isolated change on your global branch. If you’re referring to viewing a dataset version, you can select master, or view any branched versions.
2. Selecting the new dataset in Ontology Manager
Ontology Manager supports loading datasource schemas from the dataset branch associated with the selected global branch. Check that:
Ontology Manager and the transform are using the same global branch.
The output has a successful build and a readable schema on that branch.
You have access to the dataset, and the picker’s search/filter settings aren’t excluding it.
Try locating the dataset by its RID or browsing directly to its folder.
3. Workshop autosave
Try to save the module manually once on the selected branch. I believe autosave is disabled until an initial module version exists on that branch. Subsequent edits can be autosaved as drafts, but that is separate from explicitly saving a version. See Branching Workshop modules.
4. Actions created on main
If the action was created after your ontology branch diverged, try Ontology Manager → Main branch updates → Rebase branch rather than duplicating it. Once you modify the ontology on a branch—including indexing an object type—new changes from main require a manual rebase. See Ontology rebasing.
To execute the action on the branch, all object types it edits must also be indexed there. Action-generated object edits are for testing and are not merged into main. See Branching action types.