Randomly select from numpy array

Question

I have two related numpy arrays, X and y. I need to select n random rows from X and store this in an array, the corresponding y value and the appends to it the index of the points randomly selected.

I have another array index which stores a list of index which I dont want to sample.

How can I do this?

Sample data:

index = [2,3]
X = np.array([[0.3,0.7],[0.5,0.5] ,[0.2,0.8], [0.1,0.9]])
y = np.array([[0], [1], [0], [1]])

If these X's were randomly selected (where n=2):

randomylSelected = np.array([[0.3,0.7],[0.5,0.5]])

the desired output would be:

index = [0,1,2,3]
randomlySelectedY = [0,1]

How can I do this?

To be created by randomly selecting n rows from X. @Divakar — scutnex
– scutnex, Commented Apr 19, 2017 at 22:00
Could you please clarify how your index changes from [2,3] to [0,1] when it's not sampled? What's the purpose of the index and how does it relate to the other arrays? — MSeifert
– MSeifert, Commented Apr 19, 2017 at 22:26
@MSeifert index contains a list of items already sampled which should not be sampled again. — scutnex
– scutnex, Commented Apr 19, 2017 at 23:21
@scutnex In that case: Thanks for the clarification but you should rather ask a new question instead of changing the question (after it received answers) in such fundamental ways. Could you please rollback your question to the original version and ask a new question? — MSeifert
– MSeifert, Commented Apr 20, 2017 at 8:35

MSeifert · Accepted Answer · 2020-12-27 13:52:47Z

87

You can create random indices with np.random.choice:

n = 2  # for 2 random indices
index = np.random.choice(X.shape[0], n, replace=False)

Then you just need to index your arrays with the result:

x_random = X[index]
y_random = Y[index]

edited Dec 27, 2020 at 13:52

answered Apr 19, 2017 at 22:02

MSeifert

154k41 gold badges356 silver badges377 bronze badges

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

Comments

Timbus Calin · Accepted Answer · 2020-10-21 07:42:33Z

12

just to wrap @MSeifert 's answer in a function:

def random_sample(arr: numpy.array, size: int = 1) -> numpy.array:
    return arr[np.random.choice(len(arr), size=size, replace=False)]

useage:

randomly_selected_y = random_sample(Y)

edited Oct 21, 2020 at 7:42

Timbus Calin

15.2k6 gold badges49 silver badges69 bronze badges

answered May 22, 2020 at 10:33

Alon G

3,42133 silver badges31 bronze badges

1 Comment

simulate Over a year ago

Seriously why is there no function for this, that is such a common use case. So annoying.

Collectives™ on Stack Overflow

Randomly select from numpy array

2 Answers 2

Comments

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related