[FEATURE] Check completeness of hierarchies during import
srcds opened this issue · 1 comments
srcds commented
When importing generalization hierarchies for multiple attributes it can easily happen to confuse different attributes and hierarchy files.
Example:
- Start ARX and open the "example.deid" project provided as an example in the "data" folder (https://github.com/arx-deidentifier/arx/blob/master/data/example.deid)
- Click on the column “occupation” and click on “File -> Import hierarchy”
- Now we simulate a user error by selecting the hierarchy file “adult_hierarchy_workclass.csv” (https://github.com/arx-deidentifier/arx/blob/master/data/adult_hierarchy_workclass.csv) instead of the hierarchy for occupation (I think it can happen quite fast to confuse them)
- Complete the import
- When you try to anonymize the dataset the following message will appear:
Now the user realizes that he/she might have made a mistake during importing the hierarchy. However, now, the old hierarchy for the attribute is already overwritten and lost.
It would be nice if ARX displays a Warning when you want to import an incomplete hierarchy during the import.
E.g.:
Attention
The hierarchy you are about to import does not cover the full range of the attribute "occupation": hierarchy does not contain a transformation rule for value "Sales"
Continue with import?
[Yes] [No]
This request is related to #347
prasser commented
Implemented. Closing.