BUG: pandas 1.3.0 short series of type pandas.Int32DType() fails to calculate 0.75 quantile

[anarkiwi]

numpy uint32 is fine, but pd.Uint32Dtype() is not, on the same series:

>>> df = pd.DataFrame([{'x': 60}, {'x': 70}, {'x': 1514}, {'x': 1514}, {'x': 1514}, {'x': 1019}, {'x': 60}, {'x': 77}, {'x': 60}, {'x': 60}, {'x': 60}], dtype=np.uint32)
>>> df['x'].quantile(0.75)
1266.5

Now Pandas:

>>> df = pd.DataFrame([{'x': 60}, {'x': 70}, {'x': 1514}, {'x': 1514}, {'x': 1514}, {'x': 1019}, {'x': 60}, {'x': 77}, {'x': 60}, {'x': 60}, {'x': 60}], dtype=pd.UInt32Dtype())
>>> df['x'].quantile(0.75)
Traceback (most recent call last):
  File "/usr/local/lib/python3.9/site-packages/pandas/core/arrays/integer.py", line 126, in safe_cast
    return values.astype(dtype, casting="safe", copy=copy)
TypeError: Cannot cast array data from dtype('O') to dtype('uint32') according to the rule 'safe'

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python3.9/site-packages/pandas/core/series.py", line 2447, in quantile
    result = df.quantile(q=q, interpolation=interpolation, numeric_only=False)
  File "/usr/local/lib/python3.9/site-packages/pandas/core/frame.py", line 10290, in quantile
    res = self.quantile(
  File "/usr/local/lib/python3.9/site-packages/pandas/core/frame.py", line 10309, in quantile
    res = data._mgr.quantile(qs=q, axis=1, interpolation=interpolation)
  File "/usr/local/lib/python3.9/site-packages/pandas/core/internals/managers.py", line 1338, in quantile
    blocks = [
  File "/usr/local/lib/python3.9/site-packages/pandas/core/internals/managers.py", line 1339, in <listcomp>
    blk.quantile(axis=axis, qs=qs, interpolation=interpolation)
  File "/usr/local/lib/python3.9/site-packages/pandas/core/internals/blocks.py", line 1324, in quantile
    result = quantile_compat(self.values, np.asarray(qs._values), interpolation)
  File "/usr/local/lib/python3.9/site-packages/pandas/core/array_algos/quantile.py", line 46, in quantile_compat
    out = _quantile_ea_fallback(values, qs, interpolation)
  File "/usr/local/lib/python3.9/site-packages/pandas/core/array_algos/quantile.py", line 184, in _quantile_ea_fallback
    out = type(values)._from_sequence(res, dtype=values.dtype)
  File "/usr/local/lib/python3.9/site-packages/pandas/core/arrays/integer.py", line 323, in _from_sequence
    values, mask = coerce_to_array(scalars, dtype=dtype, copy=copy)
  File "/usr/local/lib/python3.9/site-packages/pandas/core/arrays/integer.py", line 231, in coerce_to_array
    values = safe_cast(values, dtype, copy=False)
  File "/usr/local/lib/python3.9/site-packages/pandas/core/arrays/integer.py", line 133, in safe_cast
    raise TypeError(
TypeError: cannot safely cast non-equivalent object to uint32

root@572eb14c4d4f:/tmp# pip3 show numpy
Name: numpy
Version: 1.21.0
Summary: NumPy is the fundamental package for array computing with Python.
Home-page: https://www.numpy.org
Author: Travis E. Oliphant et al.
Author-email: None
License: BSD
Location: /usr/local/lib/python3.9/site-packages
Requires: 
Required-by: scipy, scikit-learn, pandas
root@572eb14c4d4f:/tmp# pip3 show pandas
Name: pandas
Version: 1.3.0
Summary: Powerful data structures for data analysis, time series, and statistics
Home-page: https://pandas.pydata.org
Author: The Pandas Development Team
Author-email: [email protected]
License: BSD-3-Clause
Location: /usr/local/lib/python3.9/site-packages
Requires: pytz, numpy, python-dateutil
root@572eb14c4d4f:/tmp# python3 --version
Python 3.9.6

[phofl]

This looks like a duplicate of #42626?

[simonjayhawkins]

This looks like a duplicate of #42626?

closing