Tag columns as "intentionally not mapped" in ontology-management

[gdf]

In ontology object-type editor it would be nice to have the possibility to define that a column is intentionally not mapped.

Intention is to helps developpers to communicate together and to not map a column that is usefull in the backing dataset but not in the object-type.

It will helps to avoid this case:

• maintain a separate documentation that indicate the list of columns that are intentionally not mapped.
• map such column by error, as it is the most common habit.

What do you think about this proposal?

[bparruck]

Can you help me understand why this case would be useful? Why not just map the column anyway, but not use it?

[bkaplan]

@bparruck there would be cost implications to index the column into the ontology.

@gdf I would think to naturally drop this column from the dataset if it is not being used. Is there a reason to keep this column in the backing dataset?

Alternatively, you could rename the column in the backing dataset (i.e. not_mapped_<column_name>) or use the column description (see image) to flag this information about the column.

[gdf]

This is true that this feature request is not the most usefull. I wanted to test this idea with the community. I hope my other will be better.

I add this concerns because Ontology accept to not map some column so we can not avoid that developer will think like us one day and choose to not map every thing.

@bparruck
My use case is we have more features in dataset for deaper analysis usage (e.g. in Contour) and we think not necessary to expose to end users in Ontology.
But we could think differently and having one parent dataset with all features and not synced to Ontology + a children dataset with less feature and synced to Ontology.

@ bkaplan
Cost aspect is also to be considered that I hadn’t thought about.

Your work around is acceptable, I keep it in mind.