With https://github.com/dotnet/machinelearning/pull/5712/files, DataFrames can now be constructed from IDataViews. However, IDataView supports vector types while DataFrame doesn't, so this will throw an exception. We should add support for vector types. 1 thing to consider here is that Arrow also supports vector types, so the backing memory store for a vector type column is already defined. (ArrowStringDataFrameColumn is the closest reference for how the backing store ought to be implemented)
With https://github.com/dotnet/machinelearning/pull/5712/files, DataFrames can now be constructed from
IDataViews. However, IDataView supports vector types while DataFrame doesn't, so this will throw an exception. We should add support for vector types. 1 thing to consider here is that Arrow also supports vector types, so the backing memory store for a vector type column is already defined. (ArrowStringDataFrameColumn is the closest reference for how the backing store ought to be implemented)