ENH: MultiIndex.dtypes

[jreback]

create a mi

In [13]: mi = pd.MultiIndex.from_product([[1,2,3], list('abc'), pd.date_range('20200101', periods=2, tz='UTC')], names=['int', 'string', 'dt'])                            

You can already get .dtypes, but is slightly cumbersome. I would propose adding Multidex.dtypes (we already have MultiIndex.dtype but its always object). I think this is worth the convenience api.

In [14]: mi.to_frame().dtypes                                                                                                                                              
Out[14]: 
int                     int64
string                 object
dt        datetime64[ns, UTC]
dtype: object

# we would actually implement like this.
In [15]: pd.Series({l.name: l.dtype for l in mi.levels})                                                                                                                   
Out[15]: 
int                     int64
string                 object
dt        datetime64[ns, UTC]
dtype: object

[abdur-n-tr]

Hi, @jreback I could give a try to implement MultiIndex.dtypes, if that's okay ?

[jreback]

there is already a PR out for this

[abdur-n-tr]

I filtered the issues with no-assignee and this also came under that tag. I am a newbie here so anyone can guide how I can find the issues that need to be resolved yet.

[Aloqeely]

(we already have MultiIndex.dtype but its always object)

May I ask why is that the case? Why does it not go for the "smallest" dtype that covers all the dtypes in the multiindex, so if it has dtypes: int, float then mi.dtype returns float?

If there's no specific reason then I'll open an enhancement request