Closed willeppy closed 1 year ago
Current pandas types are in https://github.com/cmudig/AutoProfiler/blob/main/src/components/data-types/pandas-data-types.ts
The easy way would be to add new strings to these arrays, if there is a way to use pandas for the type checking that might be better
Make sure works when dype set in way like df_housing.price = df_housing.price.astype(pd.Int64Dtype())
Unable to reproduce this, we seemingly have coverage for all the types in the pandas docs so closing for now
Right now we recognize all types from numpy or raw python, but if the type is something like Int64Dtype then it is not recognized.
https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.Int64Dtype.html