Can somebody explain how numpy.where and True/False works?

Question

I am new to python and numpy. I have written code to teach myself. However, I cannot comprehend how the below code produces its result.

Input

np.where([[True, False], [True, True]], 
         [[1, 2], [3, 4]], [[9, 8], [7, 6]])

Output

array([[1, 8],
       [3, 4]])

I do not understand how this result is achieved.

You apply np.where in a list that contains 3 sublists. Are you sure this is what you need? — seralouk
– seralouk, Commented Aug 2, 2019 at 16:58
help(np.where) will explain that this takes elements from x ([[1,2],[3,4]) where condition is True, and elements from y elsewhere, in this case, the False condition is returning the 8 value from y. — David Zemens
– David Zemens, Commented Aug 2, 2019 at 16:58

T. Welsh · Accepted Answer · 2019-08-02 17:02:32Z

1

What you are seeing is a combination of broadcasting and the function of the where conditional. In the numpy.where docstring it states:

Parameters: condition : array_like, bool Where True, yield x, otherwise yield y.

In your case your boolean input is (2, 2), and is followed by two arrays of shape (2, 2).

It applies:

[True False] 
[True True ]

to x:

[1, 8]
[3, 4]

resulting in:

[1, _]
[3, 4]

and since the second element is false takes from the second input y:

[9, 8]
[7, 6]

resulting in:

[_, 8]
[_, _]

Then combines to get the output you see.

answered Aug 2, 2019 at 17:02

T. Welsh

967 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Arturo Diaz Renteria Over a year ago

Thank you. I had read the documentation. Like David Zemens the example is from there I just could not understand the logic behind this specific example. Your explanation. Cleared it up.

Collectives™ on Stack Overflow

Can somebody explain how numpy.where and True/False works?

1 Answer 1

1 Comment

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Related