BUG/API: Index.equals should check input class?

[sinhrks]

Code Sample, a copy-pastable example if possible

Decide the behavior if non Index with same values is passed. Currently, there is incompatibility between dtypes.

pd.Index([1], dtype=object).equals(pd.Series([1], dtype=object))
# False

pd.Index([1]).equals(pd.Series([1]))
# True

output of pd.show_versions()

on current master

[jreback]

these have to be dtype compat as well as values

right so the top should work

[sinhrks]

@jreback OK, always return False if input is not Index. Regarding dtype, there are some tests which allows the difference.

• https://github.com/pydata/pandas/blob/master/pandas/tests/indexes/test_numeric.py#L523

You mean following must be False?

pd.Index([1, 2]).equals(pd.Index([1., 2.]))
# True

[jreback]

no what I mean is another Index is ok, but NOT a Series.