BUG: np.take cast to out argument #25588
Description
Describe the issue:
np.take(a, indices, out=out) fails if out.dtype is not exactly a.dtype. The doc says about out:
It should be of the appropriate shape and dtype.
But what is appropriate?
Reproduce the code example:
import numpy as np
a = np.arange(3).astype(np.int32)
indices = np.arange(2)
out = np.zeros_like(indices, dtype=np.int64)
np.take(a, indices, out=out)Error message:
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
File ~/github/numpy/build-install/usr/lib/python3.11/site-packages/numpy/_core/fromnumeric.py:59, in _wrapfunc(obj, method, *args, **kwds)
58 try:
---> 59 return bound(*args, **kwds)
60 except TypeError:
61 # A TypeError occurs if the object does have such a method in its
62 # class, but its signature is not identical to that of NumPy's. This
(...)
66 # Call _wrapit from within the except clause to ensure a potential
67 # exception has a traceback chain.
TypeError: Cannot cast array data from dtype('int64') to dtype('int32') according to the rule 'safe'
During handling of the above exception, another exception occurred:
TypeError Traceback (most recent call last)
Cell In[4], line 1
----> 1 np.take(a, indices, out=out)
File ~/github/numpy/build-install/usr/lib/python3.11/site-packages/numpy/_core/fromnumeric.py:192, in take(a, indices, axis, out, mode)
95 @array_function_dispatch(_take_dispatcher)
96 def take(a, indices, axis=None, out=None, mode='raise'):
97 """
98 Take elements from an array along an axis.
99
(...)
190 [5, 7]])
191 """
--> 192 return _wrapfunc(a, 'take', indices, axis=axis, out=out, mode=mode)
File ~/github/numpy/build-install/usr/lib/python3.11/site-packages/numpy/_core/fromnumeric.py:68, in _wrapfunc(obj, method, *args, **kwds)
59 return bound(*args, **kwds)
60 except TypeError:
61 # A TypeError occurs if the object does have such a method in its
62 # class, but its signature is not identical to that of NumPy's. This
(...)
66 # Call _wrapit from within the except clause to ensure a potential
67 # exception has a traceback chain.
---> 68 return _wrapit(obj, method, *args, **kwds)
File ~/github/numpy/build-install/usr/lib/python3.11/site-packages/numpy/_core/fromnumeric.py:45, in _wrapit(obj, method, *args, **kwds)
43 except AttributeError:
44 wrap = None
---> 45 result = getattr(asarray(obj), method)(*args, **kwds)
46 if wrap:
47 if not isinstance(result, mu.ndarray):
TypeError: Cannot cast array data from dtype('int64') to dtype('int32') according to the rule 'safe'Python and NumPy Versions:
'numpy_version': '2.0.0.dev0+git20240115.bbdd595'
Runtime Environment:
not important
Context for the issue:
out = np.take(a, indices) works, but passing the out arg explicitly saves intermediate memory and could be a tiny bit faster.