I've noticed that there are some regions (feature_code = 'ADM1') that are repeated (same country and region codes). For example, some regions of Canada are duplicated, there's one entry with the region name in English, and another one in French.
Shouldn't the alternate names column table be used for this?
Other cases of repeated regions are not really the same regions with different names like the case of Canada regions.
Yes this is true. They should be in the alternate names column.
The data for Canda is from http://geobase.ca/ and the toponyms in this dataset have different ids even if they refer to the same geographical entity. There are too many toponyms with exactly the same lat/lng and the same feature code (>30'000) to find duplicates manually. If you have an idea how to find the duplicates (alternate names), let me know.