Fix for issue #57268 - ENH: Preserve input start/end type in interval…

[VISWESWARAN1998]

…_range

• closes ENH: Preserve input start/end type in interval_range #57268
• Tests added and passed if fixing a bug or adding a new feature
• All code checks passed.
• Added type annotations to new arguments/methods/functions.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.

[mroeschke]

Thanks for the PR. Needs a whatsnew note in v3.0.0.rst and a unit test.

Also could you fix this for float types too?

[VISWESWARAN1998]

ew note in v3.0.0.rst and a unit test.

Also could you fix this for float types too?

Can you give me an example for float type with expected results? Thank you!

[VISWESWARAN1998]

I have added my implementations for supporting float types. But I see in base.py that pandas doesn't support float16 data types. Should I convert them to float64 or just let pandas thow Exception?

base.py:

elif dtype == np.float16:
    # float16 not supported (no indexing engine)
    raise NotImplementedError("float16 indexes are not supported")

For example:

interval_index_float = pd.interval_range(start=np.float32(0.0), end=np.float32(0.6), freq=0.5, closed="left")
print(interval_index_float)

Gives:

IntervalIndex([[0.0, 0.5)], dtype='interval[float32, left]')

But:

interval_index_float = pd.interval_range(start=np.float16(0.0), end=np.float16(0.6), freq=0.5, closed="left")
print(interval_index_float)

will raise:

NotImplementedError: float16 indexes are not supported

I have added implementation to ignore float16 like this and checked-in my changes. If you want to let it throw an exception let me know!

My implementation:

if isinstance(start, float | np.float16) or isinstance(
                end, float | np.float16
            ):
    dtype = np.dtype("float64")
else:
    dtype = start.dtype if start.dtype == end.dtype else np.dtype("float64")
breaks = np.arange(start, end + (freq * 0.1), freq, dtype=dtype)

Current implementation safely returns float64 if it finds float16, Just like current main release!

[VISWESWARAN1998]

Completed as you asked!

[VISWESWARAN1998]

Changes have been completed as you asked!

[mroeschke]

Thanks @VISWESWARAN1998

[VISWESWARAN1998]

You are most welcome @mroeschke !