first part of v2.0 changes

[Dr-Irv]

For 2.0, a few ones since I had a little time:

• The various numeric datetime attributes of DatetimeIndex (day, month, year etc.) were previously in of dtype int64, while they were int32 for arrays.DatetimeArray. They are now int32 on DatetimeIndex also
• Level dtypes on Indexes from Series.sparse.from_coo() are now of dtype int32, the same as they are on the rows/cols on a scipy sparse matrix. Previously they were of dtype int64
• Index cannot be instantiated using a float16 dtype. Previously instantiating an Index using dtype float16 resulted in a Float64Index with a float64 dtype.

[twoertwein]

could probably combine the two numpy types into type_t[np.integer] https://numpy.org/doc/stable/reference/arrays.scalars.html#scalars

[Dr-Irv]

OK, I've made that change and added tests. Note that it would be ideal to reject invalid dtype values (e.g. np.float16), but we can only do that once we enumerate all the possible dtype values for the constructor. I left in the test for np.float16 with a comment that we will have to address that later.

[twoertwein]

Thanks @Dr-Irv (sorry, i will not have much time the next two weeks to help much with the changes for 2.0)