Hi GD team! I have a question regarding a data model warning interpretation - i had an existing table that contained the orders among other things (not as a primary key). Now i have added a new table that will serve as orders table and i want to link it to the existing one so the order in the first table becomes a foreign key of the new table. I have checked the dependencies and the warning says that the following insights and dashboards that were sourcing the orders from the first table 'will be affected'. will they be dropped or it means that it will just be replaced by the new source? thanks much for clarifying!

Hi Masha, If the warning says that certain insights and dashboards “will be affected,” it means those analytical objects are dependent on the original structure (i.e., sourcing “orders” from the first table). In GoodData Cloud, when you move or convert attributes/facts, or delete them, the affected insights, metrics, and dashboards may be dropped or become unavailable. This is also mentioned in our documentation here:

Converting or Moving Attributes/Facts: Same impact as for fact or attribute deletion.

Deleting Fact, Attribute, Dataset or Label:

Metrics: Affected metrics are no longer listed in Metrics or Analytical Designer.

Visualizations: Affected visualizations are no longer listed in Analytical Designer.

Dashboards: Affected visualizations are no longer displayed and an error message “Sorry, we can’t display this insight” is displayed in their stead.

When you delete an attribute, its dashboard filter is automatically deleted.

Monitor Model Changes

So, if you proceed with the change and the warning lists specific insights/dashboards, those objects will likely be dropped or broken—they will not be automatically remapped to the new “orders” table. You will need to manually update or recreate those insights and dashboards to use the new data source. Recommendation: • Review the list of affected objects in the warning dialog. • Consider making changes in a cloned or development workspace first. After the change, update your insights and dashboards to reference the new “orders” table as needed.

got it, thanks for the clarification, Michael!

Moving attribute from original dataset to the new dataset where it becomes primary key should preserve its identifier so metrics, visualizations and dashboards should be not broken. Other option is to introduce the new attribute while keeping the original one, replace it in objects which use original attribute, and delete it afterwards (warning shall disappear after the replacement).