Repository navigation
[Python] pyarrow can read integers as dates in list but not np.array #55315
Description
Describe the bug, including details regarding any error messages, version, and platform.
>>> import pyarrow
>>> pyarrow.array([0, 1, 2], type = "date32")
<pyarrow.lib.Date32Array object at 0x705c8e941f00>
[
1970-01-01,
1970-01-02,
1970-01-03
]
>>> pyarrow.array(np.array([0, 1, 2]), type = "date32")
Traceback (most recent call last):
File "<python-input-4>", line 1, in <module>
pyarrow.array(np.array([0, 1, 2]), type = "date32")
~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "pyarrow/array.pxi", line 414, in pyarrow.lib.array
File "pyarrow/array.pxi", line 97, in pyarrow.lib._ndarray_to_array
check_status(NdarrayToArrow(pool, values, mask, from_pandas,
File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
pyarrow.lib.ArrowNotImplementedError: Unsupported cast from int64 to date32 using function cast_date32I'm on pyarrow 26.0.0 and numpy 2.5.4.
Component(s)
Python
Activity
[-]`pyarrow` can read integers as dates in `list` but not `np.array`[/-][+][Python] `pyarrow` can read integers as dates in `list` but not `np.array`[/+]mohamd764 commentedon Oct 29, 2026
I can reproduce this on pyarrow 26.0.0 / numpy 2.5.4. The two inputs go through different paths: a Python list is converted value by value, so plain ints are accepted for date32, while a numpy array is wrapped as int64 first and then cast to the requested type. The date32 cast only has an int32 -> date32 kernel (GetDate32Cast in cpp/src/arrow/compute/kernels/scalar_cast_temporal.cc), so
pa.array([0, 1, 2]).cast("date32")fails with the same error. date64 does have int64 -> date64, which is why that one works from numpy.Until that's added, either of these works with a numpy array:
Both give 1970-01-01, 1970-01-02, 1970-01-03. An int64 -> date32 cast with an overflow check (the list path already raises "Integer too large for date32" for 2**40) would probably make the two paths consistent.
Kodiologist commentedon Oct 30, 2026
It looks like casting smaller integer types (like
int16andint8) todate32also isn't supported.huong8373tt-beep commentedon Nov 1, 2026
I plan to investigate this issue and, if the source-level regression is confirmed, submit a focused fix with C++/Python regression coverage. I will first check for any existing work and validate the overflow behavior for integer-to-date32 casts.