[C++] Support scalar aggregation over scalars

[asfimport]

See discussion on #7308. Many/most would no-op (sum, mean, min, max), but maybe they should exist and not error? Maybe they're not needed, but I could see how you might invoke a function on the result of a previous aggregation or something.

Reporter: Neal Richardson / @nealrichardson
Assignee: David Li / @lidavidm

Related issues:

• [C++] C++ array kernels framework and execution buildout (umbrella issue) (is a child of)

PRs and other links:

• GitHub Pull Request #10725

_{Note: This issue was originally created as ARROW-9056. Please see the migration documentation for further details.}

[asfimport]

Wes McKinney / @wesm:
Seems reasonable. For NumPy it's a no-op except in the case where a minimum numbers of observations is required (e.g. standard deviation with ddof > 0)

In [1]: import numpy as np

In [2]: np.sum(5)
Out[2]: 5

In [3]: np.mean(5)
Out[3]: 5.0

In [4]: np.std(5)
Out[4]: 0.0

In [5]: np.std(5, ddof=1)
C:\Miniconda\envs\pyarrow-dev\lib\site-packages\numpy\core\_methods.py:217: RuntimeWarning: Degrees of freedom <= 0 for slice
 keepdims=keepdims)
C:\Miniconda\envs\pyarrow-dev\lib\site-packages\numpy\core\_methods.py:209: RuntimeWarning: invalid value encountered in double_scalars
 ret = ret.dtype.type(ret / rcount)
Out[5]: nan

so (sample) standard deviation would probably return null but the others would just return the scalar unmodifed

[asfimport]

David Li / @lidavidm:
Issue resolved by pull request 10725
#10725