Tile and Modify a Numpy Array at the same time

Question

Suppose I have a Numpy array like so:

[10, 11, 12]

I want to copy it several times to form a new array, but subtract every element by 1 each time I copy, to produce:

[[10 11 12]
 [ 9 10 11]
 [ 8  9 10]
 [ 7  8  9]
 [ 6  7  8]
 [ 5  6  7]]

This is simple with a list comprehension:

import numpy as np
cycles = 6
a = np.array([10, 11, 12])

a = np.stack([a - i for i in range(cycles)])

However, I was wondering if there's a Numpy command that does this, or a more efficient way that doesn't use a list comprehension. I'm using Python 2.7.

Divakar · Accepted Answer · 2017-07-12 18:11:26Z

4

One approach would be with broadcasting -

a - np.arange(6)[:,None]

Sample run -

In [94]: a
Out[94]: array([10, 11, 12])

In [95]: a - np.arange(6)[:,None]
Out[95]: 
array([[10, 11, 12],
       [ 9, 10, 11],
       [ 8,  9, 10],
       [ 7,  8,  9],
       [ 6,  7,  8],
       [ 5,  6,  7]])

edited Jul 12, 2017 at 18:11

answered Jul 12, 2017 at 17:10

Divakar

222k19 gold badges273 silver badges374 bronze badges

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

2 Comments

AGN Gazer Over a year ago

I think Approach #2 is incorrect. Make your test array a = np.array([10, 1111, 112]). Then sub1_rows(a, cycles = 10000) will return array([[ 10, 1111, 112], [ 9, 10, 1111], [ 8, 9, 10], ... ])

Divakar Over a year ago

@AGNGazer Ah yes, that won't work for a generic input array that's not necessarily in sequence. Thanks for pointing it out! Removed that one.

Collectives™ on Stack Overflow

Tile and Modify a Numpy Array at the same time

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related