OpenAtlas

I turned this issue into a question because the CSV import is working as intended, see more information below, but thank you for raising the topic.

Basically the CSV import is intended for situations where data may already exist in spreadsheets or similar. Importing complex datasets would require know how about (and access to) the database and will in most cases result in writing specific scripts for the situation at hand, something we offer e.g. in cooperations.

However, the CSV import via the user interface might still be useful in some cases and can be done with little knowledge by the users themselves. We are talking about e.g. a list of actors or places, but it was never meant to be a system to access every database table which isn't useful or might even be harmful, e.g.

• class hierarchy
• classes
• properties
• property_hierarchy

Data in these tables are taken directly from the CIDOC specifications. It would not make sense or could even be harmful to mess with these.

• links
• appellation
• geometries
• object_location

These database entries would all need valid identifier to other entities. In a normal use cases, e.g. when importing an actor list, the ids would be created at the import and so, not known (directly).
Another issue beside the need for valid ids would be the need for not only valid CIDOC relations but also limit these to the ones that are implemented in the application. All in all it would be a lot of effort for a rather limited use case scenario.

Anyway, although this feature was never meant to cover every aspect, in case you have a practical use case, e.g. the actual need to import a list of administrative units (not links or model data), feel free to create an issue for this.