all nulls should be inferred as generic

[majidaldo]

i don't think this should be expected behavior.

In [41]: infer_type(pd.DataFrame({'x':['', '']}), StandardSet())
Out[41]: {'x': DateTime}

In [39]: infer_type(pd.DataFrame({'x':[None, None]}), StandardSet())
Out[39]: {'x': Boolean}

[sbrugman]

@majidaldo Thank you for reporting, these cases weren't handled properly indeed.

Sequences consisting of null values (e.g. np.nan, pd.NaT, None) will be of type Generic from now on.
The sequence of strings with length 0 will be of type String (users can of course override this behaviour).

[majidaldo]

Thanks. And pd.NA?

[sbrugman]

pd.NA is also covered. The latest release (v0.6.1) contains the fixes, feel free to try.