All ordinal values are categorical but not all categorical values are ordinal (in the implementation it’s the difference between ordered=True and ordered=False).
It looks like ordered categories in pandas support arbitrary ordering (i.e. b > z > a) there’s no way to validate a “correct” ordering. The ordering is whatever the series specifies.
We can add integer to ordinal but that’s a much better problem than it first appears because it opens the world to how opinionated do we want to be about categories. Are low cardinality integers categorial (and potentially consequently ordinal)? If so what cardinality is our threshold. Are all strings categories a la R? We’ve totally punted on that question to date I think because there’s no objective answer and depends on user intent.
model_relation(tenzing_ordinal, tenzing_categorical, inferential=False),
Isn't this always the case? Check needs to make sure all values are ordered. Possibly also add integer -> ordinal
All ordinal values are categorical but not all categorical values are ordinal (in the implementation it’s the difference between ordered=True and ordered=False).
It looks like ordered categories in pandas support arbitrary ordering (i.e. b > z > a) there’s no way to validate a “correct” ordering. The ordering is whatever the series specifies.
We can add integer to ordinal but that’s a much better problem than it first appears because it opens the world to how opinionated do we want to be about categories. Are low cardinality integers categorial (and potentially consequently ordinal)? If so what cardinality is our threshold. Are all strings categories a la R? We’ve totally punted on that question to date I think because there’s no objective answer and depends on user intent.