Let DataFrame support UInt32 columns to be compatible with MapValueToKey

[oduis]

ML.NETs Transforms.Conversion.MapValueToKey() adds a UInt32 column with an auto generated integer category key.
Unfortunately the resulting dataview crashes when trying to convert to ToDataFrame(), because it does not support UInt32.

However converting to DataFrame is essential to easily display the results in a Jupyter Notebook (which works nicely in VS Code for DataFrames).

[asmirnov82]

Hello @oduis and @luisquintanilla, is this still actual? I look through the code, Dataframe do support UInt32, so it seems there is different root cause for this. Could you please provide more info or better some code to reproduce? I would like to spend some time and try to fix this.

Can it be related to #6499 issue?

[oduis]

Sorry, haven't tried any new version. I'm currently not working with ML.NET.