polyfit time coordinate treatment in 2024.10.0

[JGuetschow]

What happened?

When using xr.polyfit on a DataArray with time coordinate (datetime64) the coefficients don't match the data but the data shifted such that t_min = 0. So the coefficients obtained can not be used in xr.polyval directly.
This shift happens in _floatize_x (

xarray/xarray/core/missing.py

Line 585 in 91962d6

def _floatize_x(x, new_x):

). With xarray 2024.10.0 the problem does not exist.

What did you expect to happen?

I expected the coefficient of the polynomial to be consistent with the data such that they can be used in xr.polyval directly.

Minimal Complete Verifiable Example

import numpy as np
import xarray as xr
import pandas as pd

# with 2024.9.0 numbers are similar, with 2024.10.0 not

test_ts = xr.DataArray(
    np.linspace(6, 12, 11),
    coords={"time": pd.date_range("1956-01-01", "1966-01-01", freq="YS")},
    dims="time",
    name="test_ts",
)

time_to_eval = np.datetime64('1957-01-01')

fit = test_ts.polyfit(dim='time', deg=1, skipna=True)
value = xr.polyval(
    test_ts.coords['time'].loc[{'time': [time_to_eval]}],
    fit.polyfit_coefficients
)

print(f"computed_value: {value.data}")
print(f"expected value: {test_ts.loc[{'time': [time_to_eval]}].data}")

MVCE confirmation

• Minimal example — the example is as focused as reasonably possible to demonstrate the underlying issue in xarray.
• Complete example — the example is self-contained, including all data and the text of any traceback.
• Verifiable example — the example copy & pastes into an IPython prompt or Binder notebook, returning the result.
• New issue — a search of GitHub Issues suggests this is not a duplicate.
• Recent environment — the issue occurs with the latest version of xarray and its dependencies.

Relevant log output

# xarray 2024.9.0
computed_value: [6.60068707]
expected value: [6.6]

# xarray 2024.10.0
computed_value: [-1.79982014]
expected value: [6.6]

Anything else we need to know?

My original issue in the primap2 package is here: primap-community/primap2#293 But that's just fyi, I think I have included all necessary information here.

Environment

INSTALLED VERSIONS

commit: None
python: 3.12.3 (main, Sep 11 2024, 14:17:37) [GCC 13.2.0]
python-bits: 64
OS: Linux
OS-release: 6.8.0-47-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: en_US.UTF-8
LANG: en_US.UTF-8
LOCALE: ('en_US', 'UTF-8')
libhdf5: 1.14.4
libnetcdf: None

xarray: 2024.10.0
pandas: 2.2.3
numpy: 1.26.4
scipy: 1.14.1
netCDF4: None
pydap: None
h5netcdf: 1.4.0
h5py: 3.12.1
zarr: None
cftime: None
nc_time_axis: None
iris: None
bottleneck: None
dask: None
distributed: None
matplotlib: 3.9.2
cartopy: None
seaborn: None
numbagg: 0.8.2
fsspec: None
cupy: None
pint: 0.24.3
sparse: None
flox: None
numpy_groupies: None
setuptools: 75.4.0
pip: 24.2
conda: None
pytest: None
mypy: None
IPython: 8.29.0
sphinx: None

[keewis]

Thanks for the report!

Will this be fixed by #9691, or is this a separate issue?

[JGuetschow]

From looking at the tests added in #9691 , I think it will (one test is basically my failing example). I only looked at the open issues and could not find a fitting one. Great that it's already fixed.

[keewis]

Well, now we have one!

I'll close this as fixed by #9691, then.