While testing preprocessing pipeline, RuntimeError was raised: RuntimeError: The column "feature_samenum_col" inferred type is floating, but only string and int categorical columns are supported.
The column contains a floating repeated value, along with few errors inserted (invalid values, substrings, ..). It is analyzed as categorical column by _columns_type property.
Should we avoid to raise the error? Or we should consider these particular columns in a different way? (maybe columns with one floating repeated values are very rare because meaningless?)
While testing preprocessing pipeline, RuntimeError was raised:
RuntimeError: The column "feature_samenum_col" inferred type is floating, but only string and int categorical columns are supported.
The column contains a floating repeated value, along with few errors inserted (invalid values, substrings, ..). It is analyzed as
categorical column
by_columns_type
property. Should we avoid to raise the error? Or we should consider these particular columns in a different way? (maybe columns with one floating repeated values are very rare because meaningless?)