ERR: invalid error reporting when comparing vs. None

[ghost]

The "in" function does not work as expected in regard to a data frame. Let's say we have the following data frame :

df = pd.DataFrame(np.random.randn(6,4), columns=list('ABCD'))

We can test this data frame :

df is None
False

However we cannot test its presence in a list :

df in [None]
TypeError: Could not compare [None] with block values

We would expect to get just False in this case.

(Pandas version 0.17.1)

[jreback]

This is coming from this, while useful (should simply be using df.isnull()) to handle all dtypes.

In [3]: df==None
TypeError: Could not compare [None] with block values

is a bug and causes __nonzero__ to fail I think.

This is an inherently ambiguous operation, see docs here

In [22]: df['A'] in [None]
ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

[mroeschke]

This is raising the correct exception on master. Could use a test.

In [33]: df = pd.DataFrame(np.random.randn(6,4), columns=list('ABCD'))
    ...:
    ...:

In [34]: df in [None]
ValueError: The truth value of a DataFrame is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

In [35]: pd.__version__
Out[35]: '0.26.0.dev0+684.g953757a3e'

[eddy-geek]

This seems to have unintended side-effects -- i.e. why should this seemingly innocuous code raise ValueError? (this is a from a real use-case around providing default values for a dataframe of plot text annotations):

df_or_string = pd.DataFrame()
if df_or_string == 'foo':
    print('bar')