Consistent behavior for array.sum(axis_from_another_array) vs array[axis_from_another_array].sum(axis_name)

[gdementen]

array[axis_from_another_array] takes the labels into account while array.sum(axis_from_another_array) does an axis aggregation (and thus ignores the labels).

>>> arr = la.ndtest(3)
>>> arr
a  a0  a1  a2
    0   1   2
>>> arr2 = la.ndtest(4)
>>> arr2
a  a0  a1  a2  a3
    0   1   2   3
>>> arr2[arr.a]
a  a0  a1  a2
    0   1   2
>>> # should be 3, not 6
... arr2.sum(arr.a) 
6

[gdementen]

priority very high because yvda hit this one and this is a silent wrong result