Previously working time series resampling breaks in new version of Dask

[andersglindstrom]

Describe the issue:

Resampling over a time series used to work with dask==2024.2.1. However, with dask=2024.5.0 and dask-expr=1.1.0 it throws an exception.

Minimal Complete Verifiable Example:

import dask.dataframe as dd

df = dd.DataFrame.from_dict({
    "Date":["11/26/2017", "11/26/2017"],
    "Time":["17:00:00.067", "17:00:00.102"],
    "Volume": [403, 3]
}, npartitions=1)

df["Timestamp"] = dd.to_datetime(df.Date) + dd.to_timedelta(df.Time)

df.to_parquet("test.parquet", write_metadata_file=True)
df = dd.read_parquet("test.parquet", index="Timestamp", calculate_divisions=True)

# This resampling breaks
minutely_volume = df[["Volume"]].resample("min").sum().compute()
print(minutely_volume)

Anything else we need to know?:
With old version:

<Deprecation warning about deprecated Dask version...>
Timestamp
2017-11-26 17:00:00     406

With new version:

Traceback (most recent call last):
  File "/home/anders/src/daskbug/daskbug.py", line 14, in <module>
    minutely_volume = df[["Volume"]].resample("min").sum().compute()
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/dask_expr/_collection.py", line 475, in compute
    out = out.optimize(fuse=fuse)
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/dask_expr/_collection.py", line 590, in optimize
    return new_collection(self.expr.optimize(fuse=fuse))
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/dask_expr/_expr.py", line 94, in optimize
    return optimize(self, **kwargs)
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/dask_expr/_expr.py", line 3028, in optimize
    return optimize_until(expr, stage)
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/dask_expr/_expr.py", line 2989, in optimize_until
    expr = expr.lower_completely()
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/dask_expr/_core.py", line 436, in lower_completely
    new = expr.lower_once()
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/dask_expr/_core.py", line 393, in lower_once
    out = expr._lower()
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/dask_expr/_resample.py", line 74, in _lower
    self.frame, new_divisions=self._resample_divisions[0], force=True
  File "/usr/lib/python3.10/functools.py", line 981, in __get__
    val = self.func(instance)
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/dask_expr/_resample.py", line 64, in _resample_divisions
    return _resample_bin_and_out_divs(
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/dask/dataframe/tseries/resample.py", line 64, in _resample_bin_and_out_divs
    temp = divs.resample(rule, closed=closed, label="left").count()
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/pandas/core/generic.py", line 9771, in resample
    return get_resampler(
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/pandas/core/resample.py", line 2050, in get_resampler
    return tg._get_resampler(obj, kind=kind)
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/pandas/core/resample.py", line 2229, in _get_resampler
    _, ax, _ = self._set_grouper(obj, gpr_index=None)
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/pandas/core/resample.py", line 2529, in _set_grouper
    obj, ax, indexer = super()._set_grouper(obj, sort, gpr_index=gpr_index)
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/pandas/core/groupby/grouper.py", line 403, in _set_grouper
    indexer = self._indexer_deprecated = ax.array.argsort(
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/pandas/core/arrays/base.py", line 848, in argsort
    return nargsort(
  File "/home/anders/src/daskbug/venv2/lib/python3.10/site-packages/pandas/core/sorting.py", line 439, in nargsort
    indexer = non_nan_idx[non_nans.argsort(kind=kind)]
TypeError: '<' not supported between instances of 'Timestamp' and 'int'

Environment:

• Dask version: 2024.5.0
• Dask-expr version: 1.1.0
• Python version: 3.10.12
• Operating System: Ubuntu 22.04.3 LTS
• Install method (conda, pip, source): pip

[andersglindstrom]

Here are two requirements file, which are the output of pip freeze in the two environments; the one that works, and the one that doesn't.
pip_freeze_working.txt
pip_freeze_broken.txt

[phofl]

thanks for the report, I'll prepare a fix

[andersglindstrom]

Thanks for fixing 👍