Type conversion failure on numpy 0.1.20

[asfimport]

While I have not dug deep enough in the Arrow codebase, it seems to me that this is caused by the new numpy release: https://github.com/numpy/numpy/releases 

The issue below in fact is not observed when using numpy 0.19.*

 

 

 

>>> pandas.__version__, pa.__version__, numpy.__version__
('1.2.1', '2.0.0', '1.20.0')
>>> df = pandas.DataFrame({'a': numpy.random.randn(10), 'b': numpy.random.randn(7).tolist() + [None, pandas.NA, numpy.nan], 'c': list(range(9)) + [numpy.nan]})
>>> pa.Table.from_pandas(df)
Traceback (most recent call last):
  File "<input>", line 1, in <module>
    pa.Table.from_pandas(df)
  File "pyarrow/table.pxi", line 1394, in pyarrow.lib.Table.from_pandas
  File "/Users/carlomazzafero/.virtualenvs/arr/lib/python3.7/site-packages/pyarrow/pandas_compat.py", line 588, in dataframe_to_arrays
    for c, f in zip(columns_to_convert, convert_fields)]
  File "/Users/carlomazzafero/.virtualenvs/arr/lib/python3.7/site-packages/pyarrow/pandas_compat.py", line 588, in <listcomp>
    for c, f in zip(columns_to_convert, convert_fields)]
  File "/Users/carlomazzafero/.virtualenvs/arr/lib/python3.7/site-packages/pyarrow/pandas_compat.py", line 574, in convert_column
    raise e
  File "/Users/carlomazzafero/.virtualenvs/arr/lib/python3.7/site-packages/pyarrow/pandas_compat.py", line 568, in convert_column
    result = pa.array(col, type=type_, from_pandas=True, safe=safe)
  File "pyarrow/array.pxi", line 292, in pyarrow.lib.array
  File "pyarrow/array.pxi", line 79, in pyarrow.lib._ndarray_to_array
  File "pyarrow/array.pxi", line 67, in pyarrow.lib._ndarray_to_type
  File "pyarrow/error.pxi", line 107, in pyarrow.lib.check_status
pyarrow.lib.ArrowTypeError: ('Did not pass numpy.dtype object', 'Conversion failed for column a with type float64')

Environment: Python 3.7.4
Mac OS

Reporter: Carlo Mazzaferro
Assignee: Uwe Korn / @xhochy

Related issues:

• [Python] Avoid usage of NumPy's PyArray_DescrCheck macro (duplicates)

_{Note: This issue was originally created as ARROW-11445. Please see the migration documentation for further details.}

[asfimport]

Uwe Korn / @xhochy:
This is known issue with wheels older versions, see also the linked issue and numpy/numpy#17913. Please update your pyarrow version if you want to use the latest NumPy release.