Unit Test Results

See test report for an extended history of previous test failures. This is useful for diagnosing flaky tests.

      9 files  ±0        9 suites  ±0   3h 5m 39s ⏱️ - 7m 31s
 18 059 tests ±0   16 842 ✅ ±0   1 217 💤 ±0  0 ❌ ±0 
161 605 runs  ±0  149 488 ✅ +1  12 117 💤  - 1  0 ❌ ±0 

Results for commit f9112c2. ± Comparison against base commit 37835c4.

♻️ This comment has been updated with latest results.

This fix appear to be OK, but there are a couple new failures:

https://github.com/dask/dask/actions/runs/17585475622/job/49952187363?pr=12071#step:11:36267

FAILED dask/dataframe/tests/test_utils_dataframe.py::test_meta_nonempty - pandas.errors.Pandas4Warning: Constructing a Categorical with a dtype and values containing non-null entries not in that dtype's categories is deprecated and will raise in a future version.
FAILED dask/dataframe/tests/test_utils_dataframe.py::test_meta_nonempty_empty_categories - pandas.errors.Pandas4Warning: Constructing a Categorical with a dtype and values containing non-null entries not in that dtype's categories is deprecated and will raise in a future version.
= 2 failed, 16339 passed, 1013 skipped, 335 xfailed, 282 xpassed, 650 warnings in 1485.05s (0:24:45) =

I can address them here.

That latest warning came from something like

pd.Categorical([0, 0], categories=[])

I changed it to use pd.Categorical.from_codes. If we have categories, we know that [0, 0] are valid codes. If we don't have any categories, we can use [-1, -1].