eq and ne not consistent with other comparison operators

[cpcloud]

works

In [1]: df = DataFrame(randint(10, size=(10, 1)), columns=['a'])

In [2]: df['dates'] = date_range('20010101', periods=len(df))

In [3]: df.a == df.dates
Out[3]:
0    False
1    False
2    False
3    False
4    False
5    False
6    False
7    False
8    False
9    False
dtype: bool

doesn't work:

In [4]: df.a <= df.dates
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-4-74e89260dea7> in <module>()
----> 1 df.a <= df.dates

/home/phillip/Documents/code/py/pandas/pandas/core/series.pyc in wrapper(self, other)
    336             if len(self) != len(other):
    337                 raise ValueError('Series lengths must match to compare')
--> 338             return self._constructor(na_op(self.values, other.values),
    339                                      index=self.index, name=name)
    340         elif isinstance(other, DataFrame):  # pragma: no cover

/home/phillip/Documents/code/py/pandas/pandas/core/series.pyc in na_op(x, y)
    325                 result = lib.scalar_compare(x, y, op)
    326         else:
--> 327             result = op(x, y)
    328
    329         return result

TypeError: invalid type promotion

[jtratner]

It's a general issue with datetime comparison to integers not being consistent. Further, == is not at all the same as <=. <= implies ordering, whereas == implies 'is the object at this position the same as the object at this other position'. earlier @jreback and I were considering an idea where we would limit boolean comparisons (except for ==) to objects of the same type.

For example, what would you expect for 1 <= date(2011, 01, 5)?

[cpcloud]

@jtratner All of those things are why I raised this issue.

[jtratner]

Got it, guess I wasn't clear earlier.