-

[DanRidh]

I'm currently trying to write an element-wise product (Hadamard product) function but am facing issues when the input fed is of a weird nested format. It works fine when lets say the input is;

input_1: [1,2,3]
input_2: [4,5,6]
output: [4,10,18]

However, when there is nesting within a specific index i'm not sure how to get the expected output (flatenning and casting seem to be throwing errors, though this is probably due to a lack of deeper understanding on my end)

eg:

input_1: [[1], [2,[3]]]
input_2: [[4], [5,[6]]]
expected_output: [[4], [10,[18]]]

Any guidance would be really helpful!

[agoose77]

If I multiple the arrays as you give them, the element-wise operation just works:

>>> x = ak.Array([[1], [2,[3]]])

>>> y = ak.Array( [[4], [5,[6]]])

>>> x * y
<Array [[4], [10, [18]]] type='2 * var * union[int64, var * int64]'>

Is this not what you observe?

[agoose77]

Could you clarify the version of Awkward Array that you're running?

>>> import awkward as ak
>>> ak.__version__

[DanRidh]

@agoose77 sure thing, I'm currently using version 2.0.8

[agoose77]

Did you copy the program directly, and run it in a new session? I can't reproduce this on my end.

[DanRidh]

@agoose77 I just installed awkward through pip, opened a kernel on jupyter-notebook and imported awkward. Could it maybe be due to python version? I'm currently using v3.8.12

[agoose77]

Could you share the traceback? And, perhaps also the notebook?