Create a python pandas MultiIndex object from separate Index objects-CodePudding

Without going into detail why I'd like to do this - is it possible to create a pandas MultiIndex from separate Index objects? I'm specifically interested in somehow preserving the specific Index types in the new MultiIndex instance. Solutions including conversions to and from a dataframe / series, etc. are fine as long as the original separate Index object types (RangeIndex, Int64Index, DatetimeIndex..) survive.

CodePudding user response：

You can create a MultiIndex from multiple Indexes using pd.MultiIndex.from_arrays:

>>> x = pd.RangeIndex(0, 10)
>>> y = pd.DatetimeIndex(pd.date_range('2020-04-01', '2020-04-10'))
>>> pd.MultiIndex.from_arrays([x, y])
MultiIndex([(0, '2020-04-01'),
            (1, '2020-04-02'),
            (2, '2020-04-03'),
            (3, '2020-04-04'),
            (4, '2020-04-05'),
            (5, '2020-04-06'),
            (6, '2020-04-07'),
            (7, '2020-04-08'),
            (8, '2020-04-09'),
            (9, '2020-04-10')],
           )