CODEWITHSUNDEEP
pythonnumpy
tix3
tix3

Reputation: 1152

Numpy column wise multiplication

I have a rather large matrix (500000 * 24) as an ndarray and I want to multiply its cell with the corresponding column min. I have already done this with for loops but I keep reading that this is not the NumPy way of doing things.

Is there a proper way of doing such an operation (I might also want to substract a constant later)?

Thanks in advance

Upvotes: 1

Views: 3039

Answers (2)

HongboZhu
HongboZhu

Reputation: 4532

Would normal multiply not do?

import numpy
a = numpy.random.random((4,2))
b = a * numpy.min(a,axis=0)

Upvotes: 1

Ffisegydd
Ffisegydd

Reputation: 53678

Yes you can simply multiply your array with the minimum vector directly, an example is shown below.

import numpy as np

data = np.random.random((500000, 24))
 # This returns an array of size 500,000 that is the row of 24 values
minimum = data.min(axis=1)

data = data * minimum

If you wish to create a minimum array of size 24 (where the minimum of the 500,000 values is taken) then you would choose axis=0.

This set of slides discusses how such operations can work.

Upvotes: 2

Related Questions