Fixes to address automlbenchmark problems

[franchuterivera]

While running the automlbenchmark, I found the following problems:

• N/A (NoResultError: cannot safely cast non-equivalent float64 to int64)
• N/A (NoResultError: Encoders require their input to be uniformly strings or numbers. Got ['int', 'str'])
• NoResultError: 'exc_info' is an invalid keyword argument for print()
• NoResultError: float() argument must be a string or a number, not 'NAType'

This is linked to problems in the dataframe preprocessing. I have added more unit test to catch this problem as well as support to address these problems.