Inconsistent isin with strings

[cpcloud]

In [5]: s = Series(list('abc'))

In [6]: s.isin('a')
Out[6]:
0     True
1    False
2    False
dtype: bool

In [5]: s = Series(list('abc'))

In [6]: s.isin('a')
Out[6]:
0     True
1    False
2    False
dtype: bool

In [7]: s = Series(['aaa', 'b', 'c'])

In [8]: s.isin('aaa')
Out[8]:
0    False
1    False
2    False
dtype: bool

I would expect:

In [7]: s = Series(['aaa', 'b', 'c'])

In [8]: s.isin('aaa')
Out[8]:
0    True
1    False
2    False
dtype: bool

or a raise....but raising seems inconsistent with the fact that single strings work....

this is happening bc the iterable passed in is turned into a set which is why a single string works

[cpcloud]

simple fix is to just turn single strings in to a list

other option is to deprecate this "feature" (it isn't tested) and enforce a list-like iterable to be passed. i'm sort of on the fence about it....practically speaking single strings are useful...but from a consistency standpoint it's ... well... inconsistent with the many of pandas other funcs/methods that take iterables

somewhat related is that not allowing single strings adds a bit more complexity to the eval parser since i must deal with things like df['a == "b"']. the == op must eval in python space. ideally that is just a call to a.isin('b') but since that doesn't work for a general string it fails a bunch of my tests