I have an np.array with 3 dimension. I just want to select all n-th element from the last array.
I know that I can select it like
array[:, :, 0]
But I need to use a function, is there any numpy function to do that ?
3
-
What do you mean by you needing to use a function? What part is variable?roganjosh– roganjosh2019-07-07 20:52:38 +00:00Commented Jul 7, 2019 at 20:52
-
1I'm not using numpy through python I just play with PythonObject so I can't use [] like in python but I can call Python function so if I can select with a function that would help meMax0u– Max0u2019-07-07 20:54:34 +00:00Commented Jul 7, 2019 at 20:54
-
1You create a function which takes a numpy array as input and returns the required element.Akash Basudevan– Akash Basudevan2019-07-07 20:56:45 +00:00Commented Jul 7, 2019 at 20:56
Add a comment
|
1 Answer
The [] operator calls __getitem__ on the object it was used on. In your case, it would be the equivalent of calling
array.__getitem__((slice(None), slice(None), 0))
This is because [:] is an empty slice (i.e. slice(None)), and you are calling the [] operator with 3 value, which is equivalent to calling __getitem__ with a tuple of 3 values.
4 Comments
Max0u
It's exactly what I was searching ! instead of using [] I needed to use a function
roganjosh
Well, fair play, I didn't see this being the answer but you've clearly seen the question in a different way to me.
Will Da Silva
@roganjosh They're interfacing with Numpy/Python from another language (as evident by their comment on the question), and so they can't use Python syntax to access elements in their array. They were looking for a function they could call in-place of the Python
[] operator.roganjosh
That makes sense. You've already got my upvote but I wouldn't have snagged this approach from what was said tbh so thanks for pointing it out :)