Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Float Index Interpreted as Object dtype #5914

theandygross opened this issue Jan 13, 2014 · 4 comments · Fixed by #5918


Copy link

commented Jan 13, 2014

When I cast a FloatIndex to a numpy array, I get an array of objects. The desired behavior should be float. This is backwards incompatible with pre-FloatIndex behavior.

In [119]:   s = pd.Series([1,2,3],[1.,2.,3.])

Out[119]:  dtype('O')

The Int64Index works as expected.

In [114]:    s = pd.Series([1,2,3],[1,2,3])

Out[114]:  dtype('int64')

This comment has been minimized.

Copy link

commented Jan 13, 2014

this is the same as 0.12

a Float64Index is actually object dtype as it makes exact matching floats easier

is their a reason you think the implementation matters here?


This comment has been minimized.

Copy link
Contributor Author

commented Jan 13, 2014

I'm running into issues where the new index is breaking code... Here is a toy example:

In [171]:

s = pd.Series([1,2,3],[1.,2.,3.])
d = pd.Series(s.index, s.index)
TypeError                                 Traceback (most recent call last)
<ipython-input-171-89e600a2d902> in <module>()
      1 s = pd.Series([1,2,3],[1.,2.,3.])
      2 d = pd.Series(s.index, s.index)
----> 3 d.idxmax()

/cellar/users/agross/anaconda2/lib/python2.7/site-packages/pandas-0.13.0_120_gdd89ce4-py2.7-linux-x86_64.egg/pandas/core/series.pyc in idxmax(self, axis, out, skipna)
   1276         DataFrame.idxmax
   1277         """
-> 1278         i = nanops.nanargmax(_values_from_object(self), skipna=skipna)
   1279         if i == -1:
   1280             return pa.NA

/cellar/users/agross/anaconda2/lib/python2.7/site-packages/pandas-0.13.0_120_gdd89ce4-py2.7-linux-x86_64.egg/pandas/core/nanops.pyc in nanargmax(values, axis, skipna)
    373     """
    374     values, mask, dtype = _get_values(values, skipna, fill_value_typ='-inf',
--> 375                                       isfinite=True)
    376     result = values.argmax(axis)
    377     result = _maybe_arg_null_out(result, axis, mask, skipna)

/cellar/users/agross/anaconda2/lib/python2.7/site-packages/pandas-0.13.0_120_gdd89ce4-py2.7-linux-x86_64.egg/pandas/core/nanops.pyc in _get_values(values, skipna, fill_value, fill_value_typ, isfinite, copy)
    139     values = _values_from_object(values)
    140     if isfinite:
--> 141         mask = _isfinite(values)
    142     else:
    143         mask = isnull(values)

/cellar/users/agross/anaconda2/lib/python2.7/site-packages/pandas-0.13.0_120_gdd89ce4-py2.7-linux-x86_64.egg/pandas/core/nanops.pyc in _isfinite(values)
    172     if issubclass(values.dtype.type, (np.timedelta64, np.datetime64)):
    173         return isnull(values)
--> 174     return -np.isfinite(values)

TypeError: ufunc 'isfinite' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ''safe''

This comment has been minimized.

Copy link

commented Jan 13, 2014

ok let me look at that
prob a bug


This comment has been minimized.

Copy link

commented Jan 13, 2014

was an issue with handling object dtypes in general in idxmin, not directly related to Float64Index...but fixed anyway

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
None yet
2 participants
You can’t perform that action at this time.