[BUG] The visualization of the dataframe resulted from the call of pd.crosstab will cause an error

[leijie-wang]

The bug could be reproduced by the following codes:
df = data.cars()
result = pd.crosstab([df.Origin], df.Cylinders)
result

The problem is that we have no control over what events are logged in the result dataframe, though generally dropna will be the final event logged. Then the lux will try to visualize the implicit tab which will finally cause an error.